OECD Social, Employment and Migration Working Papers

No. 294

The demand for language

Gabriele Marconi,
skills in the European labour
Loris Vergolini,
market: Evidence from
Francesca Borgonovi
online job vacancies
https://dx.doi.org/10.1787/e1a5abe0-en
 Organisation for Economic Co-operation and Development
1
DELSA/ELSA/WD/SEM(2023)12
DELSA/ELSA/WD/SEM(2023)12

Unclassified English text only

13 June 2023
DIRECTORATE FOR EMPLOYMENT, LABOUR AND SOCIAL AFFAIRS
EMPLOYMENT, LABOUR AND SOCIAL AFFAIRS COMMITTEE

The demand for language skills in the European labour market: Evidence from online
job vacancies

OECD SOCIAL, EMPLOYMENT AND MIGRATION WORKING PAPERS No. 294

JEL Classification: J20, J24, R10

Authorised for publication by Stefano Scarpetta, Director, Directorate for Employment, Labour and
Social Affairs.

All Social, Employment and Migration Working Papers are now available through the OECD website at
https://www.oecd.org/els/social-employment-and-migration-working-papers.htm

This working paper was prepared in the framework of the OECD Skills Outlook 2023, supported by the
European Commission through the Erasmus+ programme.

Gabriele Marconi, Agence pour le développement de l'emploi (ADEM), [email protected]

Loris Vergolini, University of Bologna and FBK-IRVAPP, [email protected]
Francesca Borgonovi, [email protected]

JT03521816
OFDE

Unclassified
This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the
delimitation of international frontiers and boundaries and to the name of any territory, city or area.
2  DELSA/ELSA/WD/SEM(2023)12

OECD Social, Employment and

Migration Working Paper

www.oecd.org/els/workingpapers
OECD Working Papers should not be reported as representing the official views of the OECD or of its
member countries. The opinions expressed and arguments employed are those of the author(s).
Working Papers describe preliminary results or research in progress by the author(s) and are published to
stimulate discussion on a broad range of issues on which the OECD works. Comments on Working Papers
are welcomed, and may be sent to [email protected].
This series is designed to make available to a wider readership selected labour market, social policy and
migration studies prepared for use within the OECD. Authorship is usually collective, but principal writers
are named. The papers are generally available only in their original language – English or French – with a
summary in the other.
This document and any map included herein are without prejudice to the status of or sovereignty over any
territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city
or area.
Note by the Republic of Türkiye
The information in this document with reference to “Cyprus” relates to the southern part of the Island. There
is no single authority representing both Turkish and Greek Cypriot people on the Island. Türkiye recognises
the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the
context of the United Nations, Türkiye shall preserve its position concerning the “Cyprus issue”.
Note by all the European Union Member States of the OECD and the European Union
The Republic of Cyprus is recognised by all members of the United Nations with the exception of Türkiye.
The information in this document relates to the area under the effective control of the Government of the
Republic of Cyprus.

© OECD 2023

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at
http://www.oecd.org/termsandconditions.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12 3

Acknowledgements

This working paper was prepared within the framework of the OECD Skills Outlook 2023, supported by the
European Commission through the Erasmus+ programme. The opinions expressed and arguments
employed herein do not necessarily reflect the official views of the OECD member countries or the
European Union.
The authors would like to thank Alexandros Bitoulas, Jiri Branka, Tiago Fragoso, Vladimir Kvetan, Lars
Ludorph, Fabio Manca, El Iza Mohamedou, Anca Nagy, Joanna Napierala, Marco Paccagnella, Raquel
Paulino, Fernando Reis, Alessio Tomelleri, Jun Yu, Quentin Vidal and OECD colleagues from the Centre
for Skills for useful comments on the paper as well as Jennifer Cannon and Duniya Dedeyn for their
administrative and editorial work.

Unclassified
4  DELSA/ELSA/WD/SEM(2023)12

Abstract

This paper investigates the demand for language skills using data on online
job vacancies in 27 European Union member countries and the
United Kingdom in 2021. Evidence indicates that although Europe remains a
linguistically diverse labour market, knowing English confers unique
advantages in certain occupations. Across countries included in the
analyses, a knowledge of English was explicitly required in 22% of all
vacancies and English was the sixth most required skill overall. A knowledge
of German, Spanish, French and Mandarin Chinese was explicitly demanded
in between 1% and 2% of all vacancies. One in two positions advertised on
line for managers or professionals required some knowledge of English, on
average across European Union member countries and across OECD
countries in the sample. This compares with only one in ten positions for
skilled agricultural, forestry and fishery workers and among elementary
occupations.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12 5

Résumé

Le présent rapport analyse la demande de compétences linguistiques en

s’appuyant sur les données relatives aux offres d’emploi en ligne publiées
dans 27 pays membres de l’Union européenne et au Royaume-Uni en 2021.
Il apparaît que même si le marché du travail européen se caractérise toujours
par une grande diversité linguistique, la connaissance de l’anglais confère
des avantages spécifiques dans certaines professions. Dans l’ensemble des
pays considérés, la connaissance de l’anglais était explicitement requise
dans 22 % des offres d’emploi et, dans l’ensemble, l’anglais constituait la
sixième compétence la plus demandée. La connaissance de l’allemand, de
l’espagnol, du français et du mandarin était explicitement requise dans 1 à
2 % des offres d’emploi. En moyenne, dans les pays membres de l’Union
européenne et dans les pays de l’OCDE faisant partie de l’échantillon,
une offre d’emploi en ligne sur deux pour des postes de cadres ou de
travailleurs intellectuels exigeait une certaine maîtrise de l’anglais. À titre de
comparaison, l’anglais n’était exigé que dans un poste sur dix concernant les
ouvriers qualifiés de l’agriculture, de la sylviculture et de la pêche et les
professions élémentaires.

Unclassified
6  DELSA/ELSA/WD/SEM(2023)12

Table of contents

OECD Social, Employment and Migration Working Paper 2

Acknowledgements 3
Abstract 4
Résumé 5
1. Introduction 8
2. Data 11
3. The demand for language skills in online job vacancies 13
3.1. The distribution of the demand for language skills in online job vacancies 13
3.2. The regional distribution of the demand for language skills in online job vacancies 17

4. The demand for languages skills across occupations and their association with
other skills 25
4.1. The variation in the demand for language skills across occupations 25
4.2. What skills are demanded alongside language skills 36
4.3. Language skills and employment conditions 38

5. Conclusions 41
Annex A. Robustness analysis on the regional distribution of the demand for English 44
Annex B. Occupational differences in the demand for language skills 45
References 51

Tables
Table 2.1. Occupational profile of vacancy data with and without detailed geographical identifiers 12
Table 3.1. Online job vacancies explicitly requiring knowledge of a language in 2021, by language 14
Table 3.2. Implicit language requirements in 2021, by country 15
Table 3.3. Explicit and implicit language requirements in online job vacancies in 2021, by country 16
Table 4.1. Skills that best predict an explicit knowledge of English, German and Mandarin Chinese, by
occupation (ISCO 1-digit) category 37

Unclassified
 DELSA/ELSA/WD/SEM(2023)12 7

Table 4.2. The association between language skills and the probability that a vacancy will offer a permanent
position 39

Table A B.1. The demand for French skills among managers, professionals, technicians, and associate
professionals 45
Table A B.2. The demand for French skills among other occupations 46
Table A B.3. The demand for German skills among managers, professionals, technicians, and associate
professionals 47
Table A B.4. The demand for German skills among other occupations 48
Table A B.5. The demand for Spanish skills among managers, professionals, technicians and associate
professionals 49
Table A B.6. The demand for Spanish skills among other occupations 50

Figures
Figure 3.1. The regional distribution of online job vacancies implicitly or explicitly requiring at least some
knowledge of English in 2021, by NUTS-3 region 18
Figure 3.2. The regional distribution of online job vacancies implicitly or explicitly requiring at least some
knowledge of German in 2021, by NUTS-3 region 20
Figure 3.3. The regional distribution of online job vacancies requiring at least some knowledge of Spanish in
2021, by NUTS-3 region 21
Figure 3.4. The regional distribution of online job vacancies requiring at least some knowledge of French in
2021, by NUTS-3 region 22
Figure 3.5. The regional distribution of online job vacancies requiring at least some knowledge of Mandarin
Chinese in 2021, by NUTS-3 region 23
Figure 3.6. The regional distribution of online job vacancies explicitly requiring at least one language skill in
2021, by NUTS-3 region 24
Figure 4.1. The demand for English skills among managers, professionals, technicians, and associate
professionals 27
Figure 4.2. The demand for English skills among clerical support workers and service and sales workers 28
Figure 4.3. The demand for English skills among skilled agricultural, forestry, fishery, craft and related trades
workers 29
Figure 4.4. The demand for English skills among plants and machine operators, assemblers and elementary
occupations 30
Figure 4.5. The demand for Mandarin Chinese skills among managers, professionals, technicians, and
associate professionals 32
Figure 4.6. The demand for Mandarin Chinese skills among clerical support workers, and among service and
sales workers 33
Figure 4.7. The demand for Mandarin Chinese among skilled agricultural, forestry, fishery, craft, and related
trades workers 34
Figure 4.8. The demand for Mandarin Chinese among plant and machine operators, assemblers, and
elementary occupations 35
Figure 4.9. Heterogeneity across occupations in the association between language skills requirements and the
likelihood that a vacancy will offer a permanent position 40

Figure A A.1. The unweighted regional distribution of online job vacancies requiring at least some knowledge
of English in 2021, by NUTS-3 region 44

Unclassified
8  DELSA/ELSA/WD/SEM(2023)12

1. Introduction

1. Language is a key component of identity (Della Chiesa, Scott and Hinton, 2012[1]; Tabouret-Keller,
2017[2]). As such, language has an important social value, notably due to its role in determining whether
an individual belongs to a specific cultural group. This can confer economic advantages (or disadvantages,
for example if some stigma is associated to the social group) in terms of social capital and networking
opportunities. Language also has a transactional value: individuals, firms, and organisations use language
to communicate with one another to exchange information, goods, and services. Transaction costs
associated with trade are rapidly decreasing because of the wide availability of affordable information and
communication technology (ICT) and speed of transportation (Hummels, 2007[3]). However, language – a
medium through which trade between different agents occurs – remains an important facilitator and barrier
to both trade (Melitz, 2008[4]) and mobility (Isphording and Otten, 2014[5]; Ovchinnikova, Mol and Jones,
2022[6]). Although human translators and interpreters as well as AI-machine translators can help individuals
and firms overcome linguistic barriers (Borgonovi, Hervé and Seitz, 2023[7]), their value lies primarily when
individuals are confronted with written text whereas language skills allow individuals to easily communicate
also orally.
2. Language proficiency is a form of human capital. The returns to language proficiency in labour
markets tend to increase with the number of speakers, vary depending on contextual factors, other forms
of human capital individuals possess and the unique combination between individual and contextual
circumstances (Chiswick and Miller, 2003[8]; Chiswick and Miller, 1995[9]). Factors that shape the
development of language proficiency among individuals include availability of high-quality and affordable
formal and informal learning opportunities, the ease with which developing proficiency occurs – which is a
function of cultural and linguistic proximity (Borgonovi and Ferrara, 2020[10]; OECD, 2018[11]), the age at
which learning occurs (Chiswick and Miller, 1995[9]; Guven and Islam, 2015[12]), and the economic and
social returns associated with gaining proficiency.
3. A large literature has examined the labour market returns associated with the development of
language skills among migrant communities. Such literature has identified large returns for immigrant
communities associated with acquiring the language spoken in the host country (Bleakley and Chin,
2004[13]; Chiswick and Miller, 1995[9]; Dustmann and Fabbri, 2003[14]; Dustmann and Soest, 2001[15]).
However, the benefits of language proficiency, defined in this work as the ability to communicate effectively
in more than one language extend also to non-migrant communities. This is especially the case in Europe,
a large, linguistically diverse community of more than 500 million people and 24 official languages in which,
thanks to the European Common market, the free movement of goods, capital, services, and people is
guaranteed. According to the Council of the European Union “language competences contribute to the
mobility, employability and personal development of European citizens, in particular young people, in line
with the objectives of the Europe 2020 strategy for growth and jobs” (Council of the European Union, 2014,
p. 2[16]). Various factors influence the language that is used in the workplace. These include the language
contents of goods and services being produced, the language of the markets in which good and services
produced are distributed and used, the language of production factors (particularly technology and the
workforce), and the language of the owners or of senior management working in a company (Grin, Sfreddo
and Vaillancourt, 2011[17]; Martel and Maurais, 1994[18]; Vaillancourt, 1996[19]).

Unclassified
 DELSA/ELSA/WD/SEM(2023)12 9

4. The ability for individuals to communicate in more than one language is associated with important
individual and societal benefits. For example, multilingualism is associated with positive economic
outcomes such as increased employability, and higher wages (Araújo et al., 2015[20]; Gazzola and
Mazzacani, 2019[21]; Ginsburgh and Prieto-Rodriguez, 2011[22]; Hahm and Gazzola, 2022[23]; Isphording,
2015[24]; Liwiński, 2019[25]). Such returns could stem from the increased productivity of workers who
possess such skills or because language proficiency is used by employers to identify characteristics such
as adaptability and openness to other cultures that are valuable but are difficult to evaluate in hiring
processes (Armstrong, 2015[26]; Hahm and Gazzola, 2022[23]). Multilingual individuals are also thought to
exhibit higher linguistic capacity, superior divergent thinking skills – a facet of creativity, as well as
enhanced attentional control, working memory, metalinguistic awareness, and abstract and symbolic
representation skills. At the same time, the economic returns to multilingualism appear to be highly
heterogeneous across different groups of individuals (Della Chiesa, Scott and Hinton, 2012[1]; Gudykunst,
2003[27]). Individuals who are able to communicate in more than one language are also more likely to
display a heightened intercultural understanding and to actively participate in global issues than
monolingual individuals (Della Chiesa, Scott and Hinton, 2012[1]; Gudykunst, 2003[27]).
5. Over the last ten years, a number of studies investigated the role of language as a skill in the
labour market, exploiting the increased data availability on this topic. Previous work conducted by Antonietti
and Loi (2014[28]) combined three sources of survey and administrative data to analyse the demand for
language skills in the Italian manufacturing sector in 2004. This study identified a strong demand for
language skills in high-skilled occupations (over 50% of the vacancies for managers and professionals
required knowledge of at least one language other than Italian), but a much lower demand for low-skilled
occupations (less than 5% of “elementary occupations” such as street vendors, hand packers and cleaners
demanded language proficiency). Fabo, Beblavý and Lenaerts (2017[29]) examine the demand for language
skills using data from online job vacancies in the Czech Republic, Hungary, Poland and the
Slovak Republic. They use data from one job portal for each country and find that a substantial proportion
of the vacancies required language skills, particularly proficiency in English. Between 28% (in the
Czech Republic) and 69% (in Poland) of online vacancies referred to proficiency in English, and vacancies
specifying English as a desirable or essential skill to perform the job on average commanded higher
salaries than similar jobs without linguistic requirements.
6. This paper builds on such studies but extends them in significant ways. First, the analysis covers
a considerably wider set of countries (all European Union countries plus the United Kingdom) and
languages (with a focus on the five most demanded languages in the European labour market: English,
German, Mandarin Chinese, Spanish and French). Second, the paper employs a novel source of data, the
Web Intelligence Hub’s Online Job Advertisement (OJA) database jointly developed by Eurostat and
Cedefop (Ascheri et al., 2022[30]; Cedefop, 2019[31]). OJA data are based on online job vacancies posted
on between 300 and 400 job portals (the exact number varies over time since some portals are closed,
others emerge and others merge over time) yielding a total of over 100 million vacancies, compared to a
total of 74 000 vacancies used in Fabo, Beblavý and Lenaerts (2017[29]). Finally, in recognition of the large
variation in labour markets conditions across regions within countries, the paper presents statistics at the
regional level. The Eurostat’s Nomenclature of Territorial Units for Statistics level 3 (NUTS-3) is used
thanks to the fine-grained location classification algorithm used in the classification of online job vacancies.
This sheds new light on the economic benefits of language skills by mapping the regional distribution of
the demand for language skills in European Union member countries and the United Kingdom.
7. The paper is structured as follows: Section 2 describes the data used. Section 3 illustrates the
languages demanded by employers advertising job openings on line in European Union member countries
and the United Kingdom. It presents regional data on the percentage of vacancies posted on line requiring
English, German, French, Spanish and Mandarin Chinese, as well as data on proficiency in at least one
language that is explicitly required in the ad. Section 4 presents data on the association between the extent
to which requirements of proficiency in English, German and Mandarin Chinese vary across occupational

Unclassified
10  DELSA/ELSA/WD/SEM(2023)12

groups. In addition, this section analyses what sets of skills are demanded alongside language skills in
different occupations and whether language skills requirements are associated with a job’s contractual
conditions. Section 5 concludes and discusses implications.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  11

2. Data

8. Analyses were conducted using a novel data source jointly produced by Eurostat and Cedefop,
the Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database (i.e., the first
release containing data for the whole of 2021) (Commission, n.d.[32]). It contains information on 162 million
vacancies that were posted on line advertising jobs in European Union countries and the United Kingdom
and that were collected between July 2018 and December 2021 (53 million ads refer to 2021, the reference
year used in this paper). The dataset was constructed by systematically crawling several hundred data
sources, for example job search engines and websites of public employment services. The dataset and
the procedures used for data collection are described in Cedefop (2019[31]) and Ascheri et al. (2022[30]).
Cedefop (2022[33]) made available some summary dashboards describing the skill data.
9. The OJA dataset contains summary information derived from the text, describing the
characteristics of the jobs which were captured through web scraping. These include, for example, if
particular educational qualifications are required to be able to apply for the job, the name of the entity that
posted the ad, the type of occupation and the industry in which the job will be performed, the salary (or
salary range) that applicants can expect to receive in compensation for their work as well as the skills
demanded by prospective employers.
10. The analyses presented in this work exploit three main pieces of information extracted from the
text of each job vacancy: Skills demanded by employers from prospective applicants interested in filling
vacancies they posted on line (at the most granular level in the European Skills, Competencies,
Qualifications and Occupations (ESCO) skills classification, at which it is possible to identify language
requirements), geographical location (NUTS-3 level) and type of occupation (the International Standard
Classification of Occupations (ISCO) 3-digit level). Information on the occupation was available for 100%
of the job vacancies. However, geographical location was available at the NUTS-3 level only for 78% of
the vacancies: 22% of the vacancies did not have detailed geographical information and were discarded
from the analysis. Table 2.1 reveals how occupations with missing information on geographical information
are similar in content to those with detailed geographical information in terms of occupation. Furthermore,
11% of the vacancies were not assigned as requiring any specific skill requirement at the most granular
skill level. These were included in the analysis as not requiring any skill. This means that estimates
presented in this work on the percentage of vacancies requiring certain languages represent a lower-bound
estimate of linguistic skills requirements since vacancies without any skill requirement were included.
Because results reflect vacancies posted on line, differences across countries, regions within countries
and across occupational groups could reflect differences in the extent to which job opportunities available
in specific contexts are posted on line. Previous work conducted using online vacancy data in fact indicated
that these are more representative of labour market opportunities in contexts with high Internet penetration,
for occupations in the tech sector, and more generally for professional and managerial occupations
(Cammeraat and Squicciarini, 2021[34]).

Unclassified
12  DELSA/ELSA/WD/SEM(2023)12

Table 2.1. Occupational profile of vacancy data with and without detailed geographical identifiers
Number and percentage of online vacancies with valid and missing information on NUTS-3 region.

Percentage across all

occupations
N (Non Among non Among
Occupation N (Missing) N (Total) % Missing
missing) missing missing
Managers 12 754 297 13 051 2.28% 8.72% 9.11%
Professional 31 418 697 32 115 2.17% 21.49% 21.39%
Technicians and
associate 23 817 521 24 338 2.14% 16.29% 15.99%
professionals
Clerical support
10 390 223 10 613 2.10% 7.11% 6.84%
workers
Service and sales
16 437 357 16 794 2.13% 11.24% 10.95%
workers
Skilled agricultural,
forestry and fishery 3 872 108 3 980 2.71% 2.65% 3.31%
workers
Craft and related
18 100 387 18 487 2.09% 12.38% 11.87%
trades workers
Plant and machine
operators, and 16 764 384 17 148 2.24% 11.47% 11.78%
assemblers
Elementary
12 651 285 12 936 2.20% 8.65% 8.75%
occupations
Total 146 203 3 259 149 462 2.18% 100.00% 100.00%

Note: Occupations are identified using the ISCO 08 classification.

Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

11. Most categorical variables were classified based on what is referred to as an “ontology matching”
model (Ascheri et al., 2022[30]). In this model, each category of each variable (for example, the category
“accountant” in the occupation variable) was associated to one or more tags (e.g. “accountant”,
“accounting analyst” and “internal auditor”). When a tag referring to an occupation was found in the vacancy
text, the vacancy was classified in that category (whenever more tags were found, the model retained the
tag with the largest number of characters and selected the respective category). In the case of occupation,
vacancies that were not classified through the ontology model were classified through a machine learning
model. Overall, the estimated accuracy of this classifier was 80% (Ascheri et al., 2022[30]). For geographical
location and skills, other variables used in this model, only the ontology matching was used.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  13

3. The demand for language skills in

online job vacancies

3.1. The distribution of the demand for language skills in online job vacancies

12. Table 3.1 illustrates the number of job vacancies explicitly requiring knowledge of each of the
languages present in the OJA dataset. Table 3.1 reflects vacancies written in any language and posted in
any country, as long as the text of the job vacancy explicitly indicated that applicants for the vacancy should
possess at least some knowledge of such language to be able to perform the job.
13. English was by far the most requested language in 2021: over 11 million vacancies were posted
explicitly specifying knowledge of English, corresponding to around 22% of all vacancies that were posted
on line in European Union countries or in the United Kingdom in 2021. The importance of English as a key
skill can also be gauged by comparing the number of job vacancies mentioning English as a required skill
with the number of vacancies mentioning other skills requirements. English was the sixth most-required
skill in online job vacancies out of all the skills classified in the most detailed level of the ESCO
classification, making English one of the most widely requested transversal skills. The five skills that were
mentioned more frequently than English in 2021 in online job vacancies were: the capacity to “adapt to
change”, “work in teams”, “use a computer”, “teamwork principles” and “use Microsoft office”. English was
mentioned in more vacancies than other widely demanded skills, including the capacity to “assist
customers”, “create solutions to problem” and “communication”. This reflect results from other studies
using online job ads and focused on smaller groups of countries (Kiškytė, 2023[35]; Fabo, Beblavý and
Lenaerts, 2017[29]).
14. German was the second most requested language: around 900 000 job vacancies were posted in
2021 explicitly requiring knowledge of German. Spanish, Mandarin Chinese and French were the third,
fourth and fifth most requested languages: around 800 000 job vacancies were posted in 2021 requiring
knowledge of Spanish, 700 000 job vacancies required knowledge of Mandarin Chinese and 600 000 job
vacancies required knowledge of French.
15. Local languages or national languages spoken by relatively small communities, such as Basque,
Finnish and Welsh, are also in relatively high demand. This could be due to the need to establish economic
ties with the local communities (for example communicating effectively with local customers), or in some
cases to other employers’ preferences (Chen, 2022[36]). Some other languages, such as Arabic, Turkish or
Bihari, probably reflect international migration and trade.

Unclassified
14  DELSA/ELSA/WD/SEM(2023)12

Table 3.1. Online job vacancies explicitly requiring knowledge of a language in 2021, by language
Number (% of total online
Language Language Ads
vacancies)
English 11 383 304 (21.6%) Russian 11 310 (<0.1%)
German 892 281 (1.7%) Czech 8 382 (<0.1%)
Spanish 812 593 (1.5%) Danish 5 265 (<0.1%)
Mandarin Chinese 659 671 (1.3%) Hungarian 3 879 (<0.1%)
French 604 827 (1.1%) Greek 2 785 (<0.1%)
Basque 417 394 (0.8%) Icelandic 2 221 (<0.1%)
Dutch 374 298 (0.7%) Slovak 1 293 (<0.1%)
Arabic 244 481 (0.5%) Croatian 1 113 (<0.1%)
Finnish 235 692 (0.4%) Turkish 1 024 (<0.1%)
Italian 193 635 (0.4%) Romanian 968 (<0.1%)
Polish 78 858 (0.1%) Slovenian 921 (<0.1%)
Welsh 75 368 (0.1%) Bulgarian 843 (<0.1%)
Norwegian 70 239 (0.1%) Bihari 520 (<0.1%)
Swedish 35 318 (<0.1%) Portuguese 416 (<0.1%)
Latvian 11 603 (<0.1%) Maltese 357 (<0.1%)

Note: Languages are sorted in descending order of the number of online job vacancies explicitly requiring a language, which indicates the
number of vacancies explicitly requiring such language in the vacancy text independently of the language in which they were written and the
country in which they were posted. Un-weighted count across European Union countries and the United Kingdom. The percentage of total online
vacancies reflects the number of online job vacancies explicitly requiring a language over the total number of job vacancies posted in 2021 in
European Union countries and the United Kingdom.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

16. Ads written in a language implicitly require some knowledge of that language. For example, if an
ad in the Netherlands is written in English (or German, Spanish or French), then this ad requires some
knowledge of English (or German, Spanish or French), even if this requirement is not made explicit in that
ad. Table 3.2 shows the prevalence of these implicit requirements, i.e. the percentage of vacancies that
were posted on line that were written in English, German, Spanish and French. Results suggest that in
many countries a large number of vacancies were written in English, even countries in which English is
not one of the official languages. For example, 58% of job postings advertised on line in Luxembourg were
written in English. By contrast, vacancies were rarely written in German, Spanish and French outside of
countries in which these languages are official languages. At the same time results indicate that unless
individuals are proficient in the official languages used in a country, their labour market opportunities
remain limited. 92% of vacancies that were posted on line in Austria in 2021 were written in German and
so were 95% of vacancies posted on line in Germany. Similarly, 98% of vacancies posted on line in 2021
in France were written in French. In Spain, Belgium, Luxembourg figures for any one language are lower
because these countries have several official languages.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  15

Table 3.2. Implicit language requirements in 2021, by country

Percentage of vacancies written in German, English Spanish or French

Country Average
German English Spanish French
Austria 91.64 7.27 0.01 0.08
Belgium 1.59 20.18 0.01 13.73
Bulgaria 2.11 15.99 0.11 0.07
Croatia 0.16 31.10 0.04 0.13
Cyprus* 0.88 55.62 0.06 0.16
Czech Republic 0.15 7.27 0.03 0.01
Denmark 0.07 11.71 0.00 0.02
Estonia 0.76 18.29 0.15 2.10
EU average 0.60 9.87 0.02 0.01
Finland 0.21 18.60 0.02 0.04
France 0.11 1.82 0.03 97.79
Germany 94.97 4.74 0.01 0.04
Greece* 2.15 45.82 0.26 0.61
Hungary 1.03 19.21 0.08 0.13
Ireland 0.11 99.54 0.02 0.07
Italy 0.20 4.64 0.06 0.51
Latvia 0.66 5.60 0.00 0.00
Lithuania 0.24 13.00 0.23 0.00
Luxembourg 4.42 57.84 0.01 37.41
Malta 0.32 98.99 0.06 0.03
Netherlands 0.67 12.24 0.02 0.08
OECD average 0.72 15.34 0.11 2.46
Poland 0.02 14.79 0.01 0.00
Portugal 1.08 30.52 1.57 0.51
Romania 0.76 32.37 1.12 0.30
Slovak Republic 0.46 8.49 0.01 0.36
Slovenia 0.73 3.63 0.00 0.01
Spain 0.29 12.42 83.74 0.33
Sweden 0.03 12.40 0.01 0.01
United Kingdom 0.36 98.99 0.01 0.15

Notes: In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia,
Cyprus, Greece, and Luxembourg (Cedefop, 2019[31]).
Vacancies written in Mandarin Chinese are rare and have been excluded by the Web Intelligence Hub’s Online Job Advertisement database.
Countries are sorted alphabetically. The EU average includes countries that are member of the European Union (EU) in the sample. The OECD
average includes countries that are member of the OECD in the sample. EU and OECD averages exclude for each column, countries with
values higher than 80% in the respective language requirement (i.e. Ireland, Malta and the United Kingdom are excluded for English; Austria
and Germany are excluded for German; Spain is excluded for Spanish; and France is excluded for French).
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
16  DELSA/ELSA/WD/SEM(2023)12

17. Table 3.3 shows, for each country, the percentage of vacancies that were posted in 2021 and that
required at least some knowledge of one of the five languages that appear to be most in demand in the
combined sample of online vacancies posted in European Union countries and the United Kingdom:
English, German, Spanish, Mandarin Chinese, French. This percentage combines vacancies that explicitly
required at least some knowledge of English, German, Spanish, Mandarin Chinese or French in the
vacancy text and vacancies that were written in one of these languages (with language skills implicitly
assumed). For example, a vacancy posted in Austria would be classified as requiring at least some
knowledge of English if the vacancy was written in German but specified that applicants would need to be
able to communicate in English to be considered. In addition, it would be classified as requiring some
knowledge of English if it did not explicitly specify any language knowledge, but it was written directly in
English. It is reasonable to expect that ads that are written in English refer to jobs where this language is
to some extent used, even though they may not mention English as a specific requirement. Finally, the last
column of Table 3.3 shows the percentage of vacancies that explicitly mentioned some knowledge of at
least one language in the vacancy text (including languages other than English, German, Spanish,
Mandarin Chinese or French) without considering the language in which the vacancy text was written.
18. When excluding countries in which English is an official language (United Kingdom, Ireland and
Malta), knowledge of English was requested in 40% or more of online vacancies posted in 2021 in Belgium,
Cyprus, France, Germany, Greece, Hungary, Luxembourg, Portugal and Romania. By contrast, knowledge
of English was requested in 15% or less of online vacancies posted in 2021 in Estonia, Latvia, and
Slovenia. For German language requirements, except for Austria and Germany, knowledge of German
was requested in 18% of vacancies posted on line only in Luxembourg, followed by 7% in Belgium. Spanish
was especially requested in Croatia and Portugal, where 15% and 29% of online vacancies respectively
requested Spanish. Mandarin Chinese was requested in over 2% of online vacancies posted in 2021 in
the United Kingdom and French was requested in 20% of vacancies posted in Belgium, and 54% of
vacancies posted in Luxembourg. On average across countries in the sample around one in five vacancies
requested at least some knowledge of a language.

Table 3.3. Explicit and implicit language requirements in online job vacancies in 2021, by country
Country English German Spanish Mandarin Chinese French At least one language explicitly required
Austria 37.05 93.85 1.50 0.46 0.12 32.73
Belgium 47.07 7.40 1.05 0.15 20.26 34.29
Bulgaria 18.54 5.67 2.46 0.00 0.29 6.50
Croatia 33.95 15.18 15.01 0.00 15.10 31.68
Cyprus* 55.62 0.88 0.06 0.00 0.16 0.00
Czech Republic 26.97 4.22 0.83 0.00 0.64 23.32
Denmark 39.82 0.30 0.00 0.00 0.02 36.36
EU 34.18 3.31 2.74 0.21 4.00 20.49
Estonia 13.62 0.60 0.02 0.00 0.01 4.92
Finland 19.66 0.21 0.02 0.00 0.04 1.28
France 47.23 1.24 1.21 1.59 98.38 48.03
Germany 40.39 97.08 2.09 1.38 0.65 38.73
Greece* 50.50 2.15 0.51 0.00 0.61 6.72
Hungary 42.24 6.53 1.47 0.00 0.29 29.73
Ireland 99.54 1.39 1.79 0.88 1.30 5.11
Italy 28.78 1.22 0.58 0.51 2.13 27.28
Latvia 11.41 0.97 0.00 0.00 0.00 7.92
Lithuania 30.21 1.00 0.23 0.00 0.00 17.95
Luxembourg 66.49 17.69 7.81 0.00 53.90 29.61
Malta 98.99 0.32 0.06 0.00 0.03 0.00

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  17

Country English German Spanish Mandarin Chinese French At least one language explicitly required
Netherlands 34.36 1.70 0.75 0.27 0.92 26.96
OECD 32.83 2.85 2.37 0.35 4.04 21.61
Poland 24.35 4.31 2.02 0.20 0.84 20.41
Portugal 41.65 4.22 28.90 0.14 4.09 36.29
Romania 56.97 2.08 2.20 0.11 0.98 27.14
Slovak Republic 15.75 1.08 0.33 0.00 0.36 8.74
Slovenia 12.98 0.73 0.00 0.00 0.01 9.35
Spain 20.58 1.03 86.98 0.00 1.04 12.62
Sweden 38.35 0.52 0.22 0.05 0.15 29.44
United Kingdom 99.26 1.43 0.76 2.46 1.52 9.21

Notes: In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia,
Cyprus, Greece, and Luxembourg (Cedefop, 2019[31]).
Countries are sorted alphabetically. The EU average includes countries that are member of the European Union in the sample. The OECD
average includes countries that are member of the OECD in the sample. EU and OECD averages exclude for each column, countries with
values higher than 80% in the respective language requirement (i.e. Ireland, Malta and the United Kingdom are excluded for English; Austria
and Germany are excluded for German; Spain is excluded for Spanish; and France is excluded for French). “English”, “German”, Spanish” and
“French” include both ads explicitly requiring this language and implicitly requiring it (because they are written in that language); “At least one
language explicitly required” includes only ads that explicitly require one of the languages listed in Table 3.1. “Mandarin Chinese” includes only
vacancies that explicitly require this language.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

3.2. The regional distribution of the demand for language skills in online job
vacancies

19. Figure 3.1 shows the proportion of online job vacancies requiring English explicitly or implicitly, by
NUTS3 region. For each NUTS3 region, the figure shows the average proportion of job vacancies requiring
English across ISCO occupational categories (3-digit). The regional distribution shows a demand for
English skills that is strong across Europe, but it is especially strong in wealthier and more densely
populated regions. This reflects results from spatial regression analysis on the same dataset (Marconi and
Vergolini, 2022[37]). The simple proportion for each region, not averaged across ISCO categories, can be
found in Figure A A.1. , in Annex A. The difference between the two figures is that Figure A A.1. reports
the simple average on the pooled sample; in Figure 3.1 first, for each ISCO occupation the proportion of
online vacancies requiring English was calculated, and then the simple average across all occupations
was computed. Results presented in Figure 3.1 account for the fact that online vacancies data suffer from
selection, i.e. certain occupations are overrepresented in the data (Cammeraat and Squicciarini, 2021[34]).
Using the arithmetic mean across occupations as suggested in the literature (Ascheri et al., 2022[30]) avoids
excessive weight to overrepresented occupations. The tables underlying all the charts in this paper are
available in the GitHub repository created by (Marconi and Vergolini, 2022[37]).
20. In around one in two regions (excluding regions in countries in which English is an official
language, namely Ireland, Malta, and the United Kingdom), at least 30% of online job ads implicitly
(i.e. being written in English) or explicitly (i.e. English being required in vacancy text) required some
knowledge of English in 2021. The percentage of online job vacancies posted in 2021 outside of countries
in which English was an official language requiring at least some knowledge of English varied between
1.5% in La Gomera (Spain) and 95% in Ithaki-Kefallinia (Greece).

Unclassified
18  DELSA/ELSA/WD/SEM(2023)12

Figure 3.1. The regional distribution of online job vacancies implicitly or explicitly requiring at least
some knowledge of English in 2021, by NUTS-3 region
Percentage of all online vacancies posted on line

>0 >0.01 >0.02 >0.03 >0.04 >0.05 >0.06 >0.07 >0.08 >0.09 >0.1 >0.2 >0.3 >0.4 >0.5 >0.6 >0.7

Note: The tables underlying all the charts in this paper are available in the GitHub repository created by Marconi and Vergolini (2022[37]). Not in
the map: Região Autónoma dos Açores 0.19, Região Autónoma da Madeira 0.15, El Hierro 0.12, Fuerteventura 0.17, Gran Canaria 0.14, La
Gomera 0.02, La Palma 0.11, Lanzarote 0.09, Tenerife 0.17, Guadeloupe 0.38, Martinique 0.37, Guyane 0.37, La Réunion 0.4, Mayotte 0.36
In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus,
Greece, and Luxembourg (Cedefop, 2019[31]). Less than 100 vacancies were identified in six NUTS-3 regions and results for these six regions
are not illustrated.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

21. Figure 3.2 indicates that outside Germany and Austria, regions with a high proportion of ads
requiring German typically reside close to the border with these two countries and in countries where
German is among the official languages like Belgium and Luxembourg. In addition, German appears to be
highly demanded in regions where the hospitality and tourism sector is an important driver of economic
activity. For instance, around one quarter of all vacancies posted in Irakleio (Greece) and Burgas (Bulgaria)
in 2021 explicitly or implicitly requested at least some knowledge of German. There is only one region
(Verviers, in Belgium) in which the share of vacancies posted on line requiring at least some knowledge of

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  19

German exceeds 30%. Data presented in Figure 3.3 suggest that only in around one-third (36%) of regions
in the European Union and the United Kingdom at least 1.5% of vacancies posted on line explicitly or
implicitly required at least some knowledge of German. For comparison, in 2021 in all regions in the
European Union and the United Kingdom at least 1.5% of online vacancies required at least some
knowledge of English.
22. Figure 3.2 exemplifies a pattern common to results on the demand for languages skills in Europe
with the exception of English: the demand for language skills is low except for few selected regions outside
of regions in which a particular language is not the official language or one of the official languages used.
Regions in which tourism and hospitality are drivers of economic growth or regions that are geographically
close to the country in which a language is the or one of the official languages used for communication
and hence large-scale local trade occurs, are the only regions in which language skills appear to be
especially valued. For example, Figure 3.3 and Figure 3.4 suggest that Spanish and French are widely
demanded in the countries in which these languages are official languages (Spain for Spanish, France,
Belgium and Luxembourg for French), in regions close to these countries or regions with important tourism
and hospitality sectors (such as the coastal area of Croatia for both Spanish and French). Spanish is also
widely requested in almost all regions in Portugal, in a specific province in Romania (Bistrita-Nasaud) and
in Zante. French is highly demanded in Bernkastel-Wittlich (a German province close to Luxembourg),
Gottingen and in Lasithi (one of the provinces of Crete, Greece).

Unclassified
20  DELSA/ELSA/WD/SEM(2023)12

Figure 3.2. The regional distribution of online job vacancies implicitly or explicitly requiring at least
some knowledge of German in 2021, by NUTS-3 region
Proportion by NUTS-3 regions, 2021

>0 >0.01 >0.02 >0.03 >0.04 >0.05 >0.06 >0.07 >0.08 >0.09 >0.1 >0.2 >0.3 >0.4 >0.5 >0.6 >0.7

Note: The tables underlying all the charts in this paper are available in the GitHub repository created by Marconi and Vergolini (2022[37]). Not in
the map: Região Autónoma dos Açores 0.01, Região Autónoma da Madeira 0.1, El Hierro 0.01, Fuerteventura 0.01, Gran Canaria 0.01, La
Gomera 0.04, La Palma 0.01, Lanzarote 0.01, Tenerife 0.01, Guadeloupe 0.01, Martinique 0.01, Guyane 0.02, La Réunion 0.01, Mayotte 0. In
the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus,
Greece, and Luxembourg (Cedefop, 2019[31]). Less than 100 vacancies were identified in six NUTS-3 regions and results for these six regions
are not illustrated.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  21

Figure 3.3. The regional distribution of online job vacancies requiring at least some knowledge of
Spanish in 2021, by NUTS-3 region
Proportion by NUTS-3 regions, 2021

>0 >0.01 >0.02 >0.03 >0.04 >0.05 >0.06 >0.07 >0.08 >0.09 >0.1 >0.2 >0.3 >0.4 >0.5 >0.6 >0.7

Note: The tables underlying all the charts in this paper are available in the GitHub repository created by Marconi and Vergolini (2022[37]). Not in
the map: Região Autónoma dos Açores 0.16, Região Autónoma da Madeira 0.2, El Hierro 0.96, Fuerteventura 0.94, Gran Canaria 0.9, La
Gomera 0.97, La Palma 0.94, Lanzarote 0.95, Tenerife 0.87, Guadeloupe 0, Martinique 0, Guyane 0.01, La Réunion 0, Mayotte 0. In the phase
of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus, Greece, and
Luxembourg (Cedefop, 2019[31]). Less than 100 vacancies were identified in six NUTS-3 regions and results for these six regions are not
illustrated.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
22  DELSA/ELSA/WD/SEM(2023)12

Figure 3.4. The regional distribution of online job vacancies requiring at least some knowledge of
French in 2021, by NUTS-3 region
Proportion by NUTS-3 regions, 2021

>0 >0.01 >0.02 >0.03 >0.04 >0.05 >0.06 >0.07 >0.08 >0.09 >0.1 >0.2 >0.3 >0.4 >0.5 >0.6 >0.7

Note: The tables underlying all the charts in this paper are available in the GitHub repository created by Marconi and Vergolini (2022[37]). Not in
the map: Região Autónoma dos Açores 0.01, Região Autónoma da Madeira 0, El Hierro 0, Fuerteventura 0, Gran Canaria 0, La Gomera 0.01,
La Palma 0, Lanzarote 0, Tenerife 0, Guadeloupe 0.99, Martinique 0.97, Guyane 0.96, La Réunion 0.93, Mayotte 0.94. In the phase of
development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus, Greece, and
Luxembourg (Cedefop, 2019[31]). Less than 100 vacancies were identified in six NUTS-3 regions and results for these six regions are not
illustrated.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

23. The demand for Mandarin Chinese (Figure 3.5) follows a different pattern from the one observed
for large European languages such as French, Spanish and German, for the simple reason that Mandarin
Chinese is not an official language in any European country. The percentage of online job vacancies that
were posted in 2021 requiring at least some knowledge of Mandarin Chinese is equal or larger than 1.5%
in about one quarter (26%) of European regions. This proportion ranges from 0% in various European
regions to 10% in Hounslow and Richmond upon Thames (United Kingdom). Among European countries,
it was highest in France, Germany and the United Kingdom, countries in which the overall demand for

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  23

workers with at least some knowledge of Mandarin Chinese exceeds 1%, on average across all regions in
the country.

Figure 3.5. The regional distribution of online job vacancies requiring at least some knowledge of
Mandarin Chinese in 2021, by NUTS-3 region
Proportion by NUTS-3 regions, 2021

>0 >0.01 >0.02 >0.03 >0.04 >0.05 >0.06 >0.07 >0.08 >0.09 >0.1 >0.2 >0.3 >0.4 >0.5 >0.6 >0.7

Note: The tables underlying all the charts in this paper are available in the GitHub repository created by Marconi and Vergolini (2022[37]). Not in
the map: Região Autónoma dos Açores 0, Região Autónoma da Madeira 0, El Hierro 0, Fuerteventura 0, Gran Canaria 0, La Gomera 0, La
Palma 0, Lanzarote 0, Tenerife 0, Guadeloupe 0.01, Martinique 0.01, Guyane 0.02, La Réunion 0.01, Mayotte 0.01. In the phase of development
of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus, Greece, and Luxembourg
(Cedefop, 2019[31]). Less than 100 vacancies were identified in six NUTS-3 regions and results for these six regions are not illustrated.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

24. Figure 3.6 shows the proportion of online job vacancies explicitly requiring at least one language
in European Union regions and regions in the United Kingdom. The proportion of online job ads explicitly
requiring at least one language exceeds 30% in almost one third (30%) of European regions, and it
exceeds 1.5% in almost all regions (96%). This proportion ranges from close to 0% in various European
regions to 67% in the Isle of Anglesey (United Kingdom). These figures are slightly lower than estimates

Unclassified
24  DELSA/ELSA/WD/SEM(2023)12

for the demand for English alone based on Figure 3.1. This is possible because Figure 3.1 includes online
vacancies written in English in addition to vacancies explicitly requiring it, while Figure 3.6 only includes
ads explicitly requiring at least one language. An important limitation of analyses of the demand for
language skills based on online job vacancies is that in certain contexts language skills are so pervasive
in a population that they can be assumed. This is the case in multilingual societies such as Malta.

Figure 3.6. The regional distribution of online job vacancies explicitly requiring at least one
language skill in 2021, by NUTS-3 region
Proportion by NUTS-3 regions, 2021

>0 >0.01 >0.02 >0.03 >0.04 >0.05 >0.06 >0.07 >0.08 >0.09 >0.1 >0.2 >0.3 >0.4 >0.5 >0.6 >0.7

Note: The tables underlying all the charts in this paper are available in the GitHub repository created by Marconi and Vergolini (2022[37]). Not in
the map: Região Autónoma dos Açores 0.2, Região Autónoma da Madeira 0.24, El Hierro 0.08, Fuerteventura 0.13, Gran Canaria 0.06, La
Gomera 0.06, La Palma 0.08, Lanzarote 0.05, Tenerife 0.08, Guadeloupe 0.39, Martinique 0.35, Guyane 0.36, La Réunion 0.38, Mayotte 0.34.
In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus,
Greece, and Luxembourg (Cedefop, 2019[31]). Less than 100 vacancies were identified in six NUTS-3 regions and results for these six regions
are not illustrated.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  25

4. The demand for languages skills

across occupations and their association
with other skills

4.1. The variation in the demand for language skills across occupations

25. This section considers how the demand for language skills differs across the following
occupations: managers; professionals; technicians and associate professionals; clerical support workers;
service and sales workers; skilled agricultural, forestry and fishery workers; craft and related trades
workers; plant and machine operators, and assemblers; elementary occupations. These occupations
correspond to broad occupational groups classified by the International Labour Organisation (ILO) using
the ISCO 1 digit level classification (ILO, 2012[38]). Workers employed in the same broad occupational
category tend to perform tasks and duties and to have a set of skills that are more similar than those of
workers employed in other occupational categories. It has to be stressed that whereas for non-manual
occupations there is a very high probability that vacancies will be posted on line, the same might not true
for manual occupations. To the extent that vacancies posted on line are more likely to also require
language skills than those not posted on line, it is possible that the demand for language skills for certain
occupations estimated using online vacancies might be higher than the demand in all vacancies for
occupations with a lower likelihood of vacancies being posted on line.
26. Analyses of differences in the demand for language skills across occupations were conducted for
English, Mandarin Chinese, German, French and Spanish. Estimates reported in Section 3 indicate that in
2021 English was the language in highest demand and the literature indicates that English is the most
widely used language, with the vast majority of users being non-native users: out of roughly
1.5 billion users of English worldwide, English was the native language only for 375 million people
(Eberhard et al., 2023[39]). English is also the means of communication of the most innovative sectors of
the economy, including the scientific community (Montgomery, 2013[40]) and the tech sector (Spence and
Liu, 2013[41]). German, French and Spanish were the other languages with a relatively large demand in
2021 in Europe. Finally, Mandarin Chinese is a language with a distinct demand pattern across Europe,
possibly linked to international trade and migration patterns. The People’s Republic of China (hereafter
‘China’) is a major exporter worldwide and Mandarin Chinese is globally the language that is spoken as a
first language by the largest number of individuals: estimates suggest that out of the 1.12 billion users of
Mandarin Chinese, only around 200 000 are non-native users of the language (Eberhard et al., 2023[39]).
27. Figure 4.1. shows the percentage of online vacancies implicitly or explicitly requiring English
language skills in managerial, professionals and technicians and associate professional occupations.
Workers in these occupations often completed tertiary or upper secondary degrees and have high levels
of foundation skills, technical skills and transversal skills to be able to complete challenging cognitive tasks
(ILO, 2012[38]). On average across EU member countries, one in two vacancies advertising managerial
and professional occupations either explicitly or implicitly require applicants to have at least some
knowledge of English. In Luxembourg around 8 in 10 vacancies advertised on line for managerial or

Unclassified
26  DELSA/ELSA/WD/SEM(2023)12

professional jobs explicitly or implicitly demanded at least some knowledge of English. English was also
highly demanded in Belgium, France, Greece, Portugal and Romania: in all these countries over 6 in 10
vacancies posted on line for managers explicitly or implicitly demanded at least some knowledge of
English. By contrast in Slovenia less than 2 in 10 vacancies posted on line for managers explicitly or
implicitly demanded at least some knowledge of English. Panel C of Figure 4.1. suggests that on average
across EU countries around 33% and 36% respectively of vacancies posted on line for technicians or
associate professionals explicitly or implicitly demanded at least some knowledge of English with Croatia
being the country with the highest demand (67%) and Slovenia the country with the weakest demand
(11%).
28. Figure 4.2 shows the percentage of online vacancies implicitly or explicitly requiring English
language skills in clerical occupations and in sales and service occupations. Panel A in Figure 4.2 suggests
that on average across OECD and EU countries 36% and 37% respectively of online vacancies for clerical
jobs implicitly or explicitly requiring English language skills. In Cyprus, France, Luxembourg and Romania,
more than 60% of online vacancies posted on line for clerical jobs implicitly or explicitly requiring English
language skills whereas in Bulgaria, Latvia, Poland, the Slovak Republic, Estonia and Finland less than
20% of online vacancies posted online for clerical jobs implicitly or explicitly requiring English language
skills. Panel B in Figure 4.2 suggests that on average across OECD and EU countries 26% respectively of
online vacancies for service and sales jobs implicitly or explicitly requiring English language skills.
29. Figure 4.3 indicates that the percentage of online vacancies for skilled agricultural forestry and
fishery workers and for craft and related trade workers implicitly or explicitly requiring English language
skills is considerably lower than the national average in all European Union countries and the
United Kingdom. This is despite the fact that vacancies for workers in these occupations that appear on
line are generally quite selected, so online postings for these occupations may be only the ones for
positions requiring greater familiarity with technology and potentially to have good communication skills,
including the ability to communicate with others in English. On average across OECD and European Union
countries only 5% and 7% respectively of online vacancies advertising positions for skilled agricultural
forestry and fishery workers implicitly or explicitly requiring English language skills. English skills are slightly
more required in vacancies looking for craft and related trade workers: on average across OECD and
European Union countries only 16% and 17% respectively of online vacancies advertising positions for
these workers implicitly or explicitly requiring English language skills.
30. The percentage of online vacancies implicitly or explicitly requiring English language skills was
lowest in vacancies advertising positions for plant and machine operators, assemblers and other
elementary occupations. Panel A in Figure 4.4 reveals that, on average, only around 3% and 2%
respectively of vacancies posted on line in OECD and European Union countries for positions for plant and
machine operators, assemblers required either implicitly or explicitly English language skills. Similarly,
Panel B in Figure 4.4 indicates that, on average, only around 6% and 7% respectively of vacancies posted
on line in OECD and European Union countries for positions for not otherwise classified elementary
occupations required either implicitly or explicitly English language skills. Results presented are in line with
results reported by Isphording (2015[24]), and Fabo, Beblavý and Lenaerts (2017[29]), and suggest that a
command of English is demanded in positions in the labour market that typically command higher wages
and require workers to solve complex analytical and cognitive tasks, like managerial and professional
technical positions.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  27

Figure 4.1. The demand for English skills among managers, professionals, technicians, and
associate professionals
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of English among
managers, professionals, technicians and associate professionals, by country
Panel A

Managers National average

%
90
80
70
60
50
40
30
20
10
0

Panel B

Professionals National average

%
90
80
70
60
50
40
30
20
10
0

Panel C

Technicians and associate professionals National average

%
90
80
70
60
50
40
30
20
10
0

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring English. These are vacancies either
specifying English as a required skill in the vacancy text or vacancies that were written in English. Data from the United Kingdom, Ireland and
Malta were excluded. “National average” refers to the average in a specific country across all occupational groups.
In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus,
Greece, and Luxembourg (Cedefop, 2019[31]).
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
28  DELSA/ELSA/WD/SEM(2023)12

Figure 4.2. The demand for English skills among clerical support workers and service and sales
workers
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of English among clerical
support workers and service and sales workers, by country

Panel A

Clerical support workers National average

%
70
60
50
40
30
20
10
0

Panel B

Service and sales workers National average

%
70
60
50
40
30
20
10
0

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring English. These are vacancies either
specifying English as a required skill in the vacancy text or vacancies that were written in English. Data from the United Kingdom, Ireland and
Malta were excluded. “National average” refers to the average in a specific country across all occupational groups.
In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus,
Greece, and Luxembourg (Cedefop, 2019[31]).
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  29

Figure 4.3. The demand for English skills among skilled agricultural, forestry, fishery, craft and
related trades workers
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of English among skilled
agricultural, forestry and fishery workers and among craft and related trades workers, by country

Panel A

Skilled agricultural, forestry and fishery workers National average

%
70
60
50
40
30
20
10
0

Panel B

Craft and related trades workers National average

%
70
60
50
40
30
20
10
0

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring English. These are vacancies either
specifying English as a required skill in the vacancy text or vacancies that were written in English. Data from the United Kingdom, Ireland and
Malta were excluded. “National average” refers to the average in a specific country across all occupational groups.
In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus,
Greece, and Luxembourg (Cedefop, 2019[31]).
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
30  DELSA/ELSA/WD/SEM(2023)12

Figure 4.4. The demand for English skills among plants and machine operators, assemblers and
elementary occupations
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of English among plants and
machine operators and assemblers and among elementary occupations, by country

Panel A

Plant and machine operators, and assemblers National average

%
70
60
50
40
30
20
10
0

Panel B

Elementary occupations National average

%
70
60
50
40
30
20
10
0

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring English. These are vacancies either
specifying English as a required skill in the vacancy text or vacancies that were written in English. Data from the United Kingdom, Ireland and
Malta were excluded. “National average” refers to the average in a specific country across all occupational groups.
In the phase of development of the database, some difficulties were encountered in guaranteeing data coverage for Bulgaria, Croatia, Cyprus,
Greece, and Luxembourg (Cedefop, 2019[31]).
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  31

31. Figure 4.5 and Figure 4.6 report the results for the percentage of online ads requiring Mandarin
Chinese. In this case, the picture is almost reversed compared to English. In fact, Mandarin Chinese is
most requested in online vacancies advertising jobs in service and sales occupations, in crafts, plant and
machine operators. This could mean that Mandarin Chinese may be requested in jobs that service primarily
Chinese migrant communities, cater to Chinese tourists travelling to Europe or be associated with
contacting trade partners in China.
32. The country with the highest demand for Mandarin Chinese in practically all the occupation groups
is the United Kingdom, followed by France, and Germany. In general, Mandarin Chinese is requested in
very few online ads: 2.5% in the United Kingdom, 1.6% in France, 1.4% in Germany and about 1% In
Ireland and Italy. These countries experienced in the last decades a huge inflow of Chinese nationals
(Plewa and Stermšek, 2017[42]), which can explain a strong demand for a knowledge of Mandarin Chinese,
especially in occupations in which large migrant communities are employed or in which contact with China
because of trade is intense. Interestingly, despite the large influx of Chinese migrants, very few online
vacancies advertised in Spain required a knowledge of Mandarin Chinese.

Unclassified
32  DELSA/ELSA/WD/SEM(2023)12

Figure 4.5. The demand for Mandarin Chinese skills among managers, professionals, technicians,
and associate professionals
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of Mandarin Chinese among
managers, professionals, technicians, and associate professionals, by country

Panel A

Managers National average

%
4
3.5
3
2.5
2
1.5
1
0.5
0

Panel B

Professionals National average

%
4
3.5
3
2.5
2
1.5
1
0.5
0

Panel C

Technicians and associate professionals National average

%
4
3.5
3
2.5
2
1.5
1
0.5
0

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring Mandarin Chinese. These are
vacancies specifying Mandarin Chinese as a required skill in the vacancy text since no vacancies were written in Mandarin Chinese.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  33

Figure 4.6. The demand for Mandarin Chinese skills among clerical support workers, and among
service and sales workers
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of Mandarin Chinese among
clerical support workers, and among service and sales workers, by country

Panel A

Clerical support workers National average

%
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0

Panel B

Service and sales workers National average

%
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring Mandarin Chinese. These are
vacancies specifying Mandarin Chinese as a required skill in the vacancy text since no vacancies were written in Mandarin Chinese.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
34  DELSA/ELSA/WD/SEM(2023)12

Figure 4.7. The demand for Mandarin Chinese among skilled agricultural, forestry, fishery, craft,
and related trades workers
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of Mandarin Chinese among
skilled agricultural, forestry and fishery workers, and among craft and related trades workers, by country

Panel A

Skilled agricultural, forestry and fishery workers National average

%
8
7
6
5
4
3
2
1
0

Panel B

Craft and related trades workers National average

%
8
7
6
5
4
3
2
1
0

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring Mandarin Chinese. These are
vacancies specifying Mandarin Chinese as a required skill in the vacancy text since no vacancies were written in Mandarin Chinese.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  35

Figure 4.8. The demand for Mandarin Chinese among plant and machine operators, assemblers,
and elementary occupations
Percentage of online vacancies implicitly or explicitly requiring at least some knowledge of Mandarin Chinese among
plant and machine operators, assemblers, and elementary occupations, by country

Panel A

Plant and machine operators, and assemblers National average

%
6

Panel B

Elementary occupations National average

%
6

Note: Countries are sorted in descending order of the number of vacancies explicitly or implicitly requiring Mandarin Chinese. These are
vacancies specifying Mandarin Chinese as a required skill in the vacancy text since no vacancies were written in Mandarin Chinese.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

33. Results for German, French and Spanish are reported in Annex B and show a pattern similar to
the one identified for English: German, French and Spanish tend to be in higher demand in managerial
and professional occupations and lower in manual occupations such as plant and machine operators,
assemblers, skilled agricultural, forestry and fishery workers. For example, German is widely required for
managerial professions in Luxembourg, Belgium and in some East-European countries (i.e. Bulgaria and
the Czech Republic), while the demand for French is very high in Luxembourg and less in other countries
with the remarkable exception of technicians in Croatia. The knowledge of Spanish is required for
managerial professions as well as for clerical, service, sales craft workers and for plant and machine
operators and assemblers. In Croatia there is a huge Spanish demand for service and sales workers.

Unclassified
36  DELSA/ELSA/WD/SEM(2023)12

4.2. What skills are demanded alongside language skills

34. Previous sections identify large differences in the demand for language skills across occupations.
In particular, online vacancies advertising positions for managers and professionals are considerably more
likely to explicitly require at least some knowledge of English and, to a lesser extent, of other European
languages such as German, French or Spanish, than other online vacancies advertising positions for other
occupational groups. By contrast, Mandarin Chinese is most demanded in service and sales occupations,
in craft and related trades occupations, and in job advertising positions for plant and machine operators
and assemblers.
35. Table 4.1 identifies, for each occupational group, which skills appear more frequently alongside
language skills in online vacancies advertising positions in each broad occupational group. This allows to
identify what other forms of human capital complement language skills (or are complemented by language
skills). For each occupation at the ISCO 1-digit level, analyses were conducted to identify among the sets
of skills included in the text of the online vacancies, which skills best predicted the explicit presence of
English, (or German, or Chinese) among the job requirements.
36. Table 4.1 relies on a widely used criterion to evaluate prediction accuracy – the percentage of
correctly predicted outcome (Poldrack, Huckins and Varoquaux, 2020[43]; Abbott, 2014[44]). It identifies, for
each language analysed – English, German and Mandarin Chinese – the “top predicting” skill and the
percentage of online vacancies that were correctly predicted as explicitly specifying language skills based
on this skill. A skill was considered to correctly predict a vacancy if its presence among (or absence from)
the job requirements of such vacancy was systematically related with the presence (or absence) of a
requirement for knowledge of a particular language in the job requirements.1 Because most vacancies did
not include any language skills requirements for a knowledge of Mandarin Chinese and German, the
percentage of correctly predicted outcomes for these languages is a valid criterion to identify top predictors,
but this indicator cannot be compared across occupations or languages.
37. Table 4.1 reveals important differences in the skills that are most strongly associated with
requirements for English, German and Mandarin Chinese. For English, the top predicting skills across the
pool of all occupational groups was “business analysis”. “Business analysis” was the top predicting skills
for an English language requirement in vacancies posted on line advertising positions for technicians and
associate professionals, clerical support workers, service and sales workers, assemblers, and for plant
and machine operators. This suggests that English is required in jobs that also demand an understanding
of business strategy and of the economic context. “Trademarks” and “Search engine optimisation” were
the top predicting skills for an English language requirement in vacancies posted on line advertising
positions for managers and professionals, signalling the commercial and technological content of positions
requiring English proficiency.
38. By contrast, Table 4.1 suggests that for German, in most occupational categories, the top
predicting skills of German linguistic requirements were other language skills (Dutch, French, Hungarian
and Spanish). This suggests that vacancies requiring German tend to refer to positions in which multiple
languages are required, for example positions requiring extensive communication with other countries
because of trade or tourism. When all vacancies were considered, the top predicting skills for a German
language requirement in vacancies posted on line was “language teaching methods”, an indication that

1
For each linguistic skill, two possible values of the percentage of correctly predicted vacancies were calculated and the highest
between the two was retained. The first was the value calculated under the assumption of a positive association between predicting
skills and target language skills (i.e. that when a certain predicting skill was included in a vacancy, target language linguistic
requirements were more likely to occur). The second was the value calculated under the assumption of a negative association
between predicting skills and the target language skills (i.e. the absence of predicting skills would be associated with the presence
of linguistic skills requirements and the presence of predicting skills would be associated with the absence of linguistic skills
requirements). In practice, top predicting skills were all positively related to linguistic skills requirements.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  37

German is an especially strong requirement in positions of German language teachers. “Language

teaching methods” was the top predicting skill for a German language requirement in vacancies for
professionals, which comprised the largest group overall in which German was required and teachers are
an especially large group of professionals.
39. Finally, Table 4.1 indicates that the sets of skills that best predicted Mandarin Chinese linguistic
requirements tended to be very occupation-specific skills: for example “control kiln firing” was the skill most
associated with Chinese in vacancies posted on line for professionals whereas “babysitting” was the skill
most associated with Chinese in vacancies posted on line for elementary occupations. For managers, the
skill that best predicted Mandarin Chinese linguistic requirements was “take food and beverage orders
from customers”, which could indicate that a relatively large part of managerial positions requiring Mandarin
Chinese involve managing small establishments (like restaurants) in which the manager is also directly
involved in service delivery.

Table 4.1. Skills that best predict an explicit knowledge of English, German and Mandarin Chinese,
by occupation (ISCO 1-digit) category
Percentage of correctly
Occupation Top predicting skill
predicted ads
English German Chinese English German Chinese
take food and
trademarks Spanish beverage orders from 78.2 97.1 99.5
Managers
customers
search engine language teaching
control kiln firing 76.9 97.7 99.2
Professionals optimisation methods

decision support
Technicians and business analysis Dutch 75.7 98.5 98.3
systems
associate professionals

business analysis Hungarian aerodynamics 71.6 98.1 99.2

Clerical support workers

business analysis Dutch climbing equipment 77.8 98.7 98.2

Service and sales workers

Skilled agricultural, precision measuring

NA aerodynamics 99.1 NA 98.8
forestry and fishery instruments
workers

Craft and related trades material mechanics legal studies ski 81.8 99.6 98.2
workers

Plant and machine

business analysis French Android 82.8 99.1 98.2
operators, and
assemblers

robotic components advocate for others babysitting 94.9 99.9 98.5

Elementary occupations
language teaching decision support
business analysis 78.6 98.3 98.7
All occupations combined methods systems

Note: In the analyses reported in this table, vacancies were defined as requiring English, German and Mandarin Chinese only if they explicitly
required that language in the vacancy text (independently of the language in which the vacancy was written). No top predicting skill for German
were found in category Skilled agricultural, forestry and fishery workers due to a limited number of vacancies requiring German in that category.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
38  DELSA/ELSA/WD/SEM(2023)12

4.3. Language skills and employment conditions

40. This section investigates if an implicit or an explicit demand for specific language skills
(i.e., English, German and Mandarin Chinese) in the online job vacancies was associated with greater
labour market returns in terms of job contract. Since the OJA database contains 162 million ads, a random
sample of vacancies that were representative at the regional level (NUTS-2) was selected to maintain the
computations manageable. In principle, this means that estimates are less precise than would have been
the case had all vacancies been used in the analysis (even though, in practice, the number of ads in the
sample is very large). The extraction process was done in two steps. In the first step, a random sample of
about 100 000 vacancies for each language was selected and, in a second step, a further random sample
of about 1.5 million vacancies was added to the final sample. The first step ensured that the sample
contained enough vacancies for each language since few vacancies required Mandarin Chinese or
German compared to English. The second step enlarged the overall sample and ensured generalisability.
In the extraction process, the condition of not having missing values on the following variables: occupation,
industry, country, region (NUTS-2) and contractual condition was imposed. This condition led to a reduction
in the population of vacancies that could be sampled: the percentage of vacancies that were excluded due
to missing information amounted to 34.3%.
41. The analytical strategy was based on a set of nested models aimed at understanding the
association between the presence of language skills requirements and contractual conditions and the
potential mediating role of occupation (ISCO 1-digit) and industry (NACE 1-digit). In order words, analyses
allow to estimate across all labour markets if vacancies containing language skills requirements are more
likely to be for permanent positions and if this association is due to the fact that permanent positions are
more likely to be offered in industries and occupations in which language skills are most required. In other
words, results presented in Table 4.2 allow to identify economy wide associations as well as the strength
of associations within occupations and industries. Table 4.2 presents three different sets of models: the
first contains as controls work experience, the language in which the vacancy was written and country fixed
effects, the second additionally controls for industry fixed effects and the last model additionally controls
for occupation fixed effects.
42. Table 4.2 shows the percentage point change in the proportion of a vacancy offering a permanent
rather than a non-permanent contract associated with explicit or implicit linguistic requirements. Online
vacancies requiring English and German are more likely to offer positions with a permanent contractual
arrangement whereas positions for which Mandarin Chinese is required are more likely to be for positions
that do not offer a permanent contract. The strength of the association between language requirements
and the likelihood that a vacancy will advertise a permanent contract is weaker when occupations as well
as industry fixed effects are included, although occupation appears to be the primary mediation
mechanism.
43. For example, a vacancy requiring English was three-percentage points more likely to offer a
permanent contract than other vacancies. However, when controlling for occupational category this
difference was just 0.6 percentage points. Similarly, the presence of German in a vacancy was associated
with a 7.5 percentage point higher probability that a vacancy advertised a position with a permanent
contract but when comparing vacancies advertised for positions with different language requirements but
for the same occupational category this difference was just 4.5 percentage points. By contrast, the
presence of Mandarin Chinese in a vacancy was associated with a four-percentage point lower probability
that a vacancy advertised a position with a permanent contract but when comparing vacancies advertised
for positions with different language requirements but for the same occupational category this difference
was just one percentage point.
44. Given these results, to deepen the discussion of the influence of the demand of English, German
and Chinese, Figure 4.9 highlights differences in the likelihood that a vacancy advertised permanent
positions as a function of language skills requirements among different occupations. Results suggests that

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  39

English skills requirements are associated with a ‘quality of contractual conditions premium’ only among
managers and professionals, whereas a German skills requirement was associated with a higher likelihood
that a vacancy advertised job opportunities with permanent contracts in almost all the occupational
categories with the exception of “Craft and related trades workers” and “Elementary occupations”.
Occupational differences in the likelihood that jobs advertised on line will be for permanent positions
depending on the presence of Mandarin Chinese skills requirements are mixed: such requirement was
associated with an increased likelihood among managers and associate professionals but a lower
likelihood among many other professional groups.

Table 4.2. The association between language skills and the probability that a vacancy will offer a
permanent position
Changes in the likelihood that a vacancy will offer a permanent position as a function of the presence in the vacancy
of a language requirement

Model 1 Model 2 Model 3

English 0.033*** 0.029*** 0.006***

(0.001) (0.001) (0.001)
German 0.075*** 0.065*** 0.045***
(0.002) (0.002) (0.002)
Mandarin Chinese -0.044*** -0.034*** -0.010***
(0.002) (0.002) (0.002)

Other controls YES YES YES

Industry NO YES YES
Occupation NO NO YES
Observations 660 183 717 062 818 030 660 183 717 062 818 030 660 183 717 062 818 030
R2 0.113 0.206 0.192 0.127 0.223 0.207 0.138 0.233 0.217

Note: Parameters derived from three sets of OLS regression models in which estimated changes in the likelihood that a vacancy offered a
permanent position were regressed on the presence of English, German and Mandarin Chinese. Data from the United Kingdom, Ireland and
Malta were excluded in Models with English. Data from Austria and Germany were excluded in Models with German. Robust standard errors in
parentheses.
* p < 0.10, ** p < 0.05, *** p < 0.01.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
40  DELSA/ELSA/WD/SEM(2023)12

Figure 4.9. Heterogeneity across occupations in the association between language skills
requirements and the likelihood that a vacancy will offer a permanent position
Predicted probabilities of having a permanent position and respective confidence intervals from OLS models
considering the interaction between foreign languages (English, German and Mandarin Chinese) and occupational
groups
English

Not required Required

0.7

0.6

0.5

0.4

0.3

0.2

0.1
Managers Professionals Technicians and Clerical support Service and sales Craft and Plant and machine Elementary
associate workers workers related trades operators, and occupations
professionals workers assemblers

German

Not required Required

0.7

0.6

0.5

0.4

0.3

0.2

0.1
Managers Professionals Technicians and Clerical support Service and sales Craft and Plant and machine Elementary
associate workers workers related trades operators, and occupations
professionals workers assemblers

Mandarin Chinese

Not required Required

0.7

0.6

0.5

0.4

0.3

0.2

0.1
Managers Professionals Technicians and Clerical support Service and sales Craft and Plant and machine Elementary
associate workers workers related trades operators, and occupations
professionals workers assemblers

Note: Parameters derived from nine sets of OLS regression models in which, for each major occupational group, estimated changes in the
likelihood that a vacancy offered a permanent position were regressed on the presence of English, German, and Mandarin Chinese. Data from
the United Kingdom, Ireland and Malta were excluded in Models with English. Data from Austria and Germany were excluded in Models with
German. Robust standard errors in parentheses.
* p < 0.10, ** p < 0.05, *** p < 0.01.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  41

5. Conclusions

45. The analyses presented in this work indicate that Europe remains a linguistically diverse labour
market: individuals who want to find employment in a specific country have to be able to master the official
language/s used in that national context. Language remains an important ‘intangible’ barrier to the
movement and allocation of workers in Europe. As a result, it is possible that some economic sectors may
suffer from an oversupply of qualified workers to perform specific tasks while other areas may face an
undersupply of such workers because workers lack the specific language skills needed to be integrated in
another linguistic context. Language learning is an important priority in the European Union, with significant
investments being dedicated to language teaching and learning in education and vocational training as
well as to promoting language learning through youth mobility.
46. Promoting language learning is not only important to promote workers’ mobility but is also an
important way for economies to reduce barriers to international trade facilitating collaboration across
linguistic boundaries. Results presented in this work indicate that on average across European Union
member countries and the United Kingdom, over one in ten vacancies that were posted on line in 2021
explicitly demanded language skills. In France and Germany this share was as high as 45% and 32%
respectively but at least some knowledge of a language other than the language in which the vacancy was
drafted was required in over 15% of all vacancies also in Austria, Belgium, Denmark, Hungary, Italy, the
Netherlands, Portugal and Sweden. Because these estimates are derived based on vacancies that were
posted on line, they reflect the demand for workers who are likely to have access to digital technologies
and use them in their job search. Such selection might lead to an overestimation of the demand for
language skills since such demand is likely to be more likely to capture workers with greater skills. At the
same time, it is possible that in many contexts in which language skills are essential to perform a job,
employers will assume that prospective workers will possess such skills and use the limited space available
in the vacancy text to spell out other essential or desirable skills. Such cases might lead to an
underestimation of the demand for language skills.
47. Despite a marked linguistic heterogeneity, analyses presented suggest that a knowledge of
English confers distinct advantages in European labour markets, especially in managerial, professional
and associate professional occupations. A knowledge of English was the sixth most demanded skill in
vacancies posted on line by employers in 2021. On average across European Union member countries
(excluding Malta and Ireland), 22% of all vacancies posted on line in 2021 either explicitly required at least
some knowledge of English or implicitly required it since the vacancy text was directly written in English.
In France this figure was almost one in two (48%) and was over one in five in Austria, Belgium, Cyprus,
Germany, Denmark, Greece, Croatia, Hungary, Luxembourg, the Netherlands, Portugal, Romania and
Sweden. English appears to be the only language that approximate a lingua franca in Europe. A knowledge
of other European languages such as French, German and Spanish or languages that are widely used in
the world such as Mandarin Chinese are not equally demanded in European labour markets. A knowledge
of German, Spanish, French and Mandarin Chinese was explicitly demanded in only between 1% and 2%
of all vacancies in European Union member countries or the United Kingdom.
48. Detailed analyses on language skills requirements in different occupations suggest that the ability
to use English was especially demanded in occupations requiring workers to perform complex cognitive
tasks such as managers and professionals and an explicit or implicit knowledge of English was most

Unclassified
42  DELSA/ELSA/WD/SEM(2023)12

demanded in occupations requiring skills related to an understanding of business strategy. A knowledge

of Mandarin Chinese was demanded especially in occupations that require less complex cognitive tasks.
In contrast, German was relatively highly demanded in vacancies for jobs that also required a knowledge
of other languages or in teaching positions, suggesting that it is demanded in jobs requiring multiple
languages (e.g. in the tourism and hospitality industry) and in the teaching of the German language.
49. In the European Union, where multilingualism is explicitly recognised as critical to realising a vision
of Europe that includes mobility, intercultural understanding and economic prosperity, young people are
generally exposed at school to a second language from primary school and an additional language in
secondary school and language education is a core component of school curricula. The most recent data
from Eurostat on language learning collected in 2015 indicates that 99% of all students in lower secondary
school were learning at least one ‘foreign’ language, defined in most countries, as the official language of
instruction (Eurostat, 2017[45]) and as many as 59% of students were learning two foreign languages or
more. English was the most studied language: 93% studied English; 34% studied French; 23% studied
German; 14% studied Spanish; 3% studied Russian and 1% studied Italian.
50. Such efforts have translated into high levels of youngsters reporting having at least some levels of
proficiency in a foreign language. In 2016, the year with the most recent available data, only around 19%
of 25- to 34-year-olds in the European Union reported that they did not know any ‘foreign language’. This
share was higher among older cohorts and was highest among individuals between the age of 55 and 64:
43% of respondents in this age group reported not knowing any ‘foreign language’. Overall, across the
European Union 32% of individuals reported not knowing any foreign language. A knowledge of foreign
languages was higher among individuals with higher levels of educational attainment, those who were
employed and, especially, among individuals working in professional, managers and associate
professional roles. For example, 82.5% of individuals in European Union countries in 2016 who had
completed a tertiary level of education reported that they knew at least one foreign language whereas only
around 42% of those who had not completed an upper secondary degree did (Eurostat, n.d.[46]).
51. As countries grapple with tight labour market conditions and a shortage of skilled workers, some
politicians and business leaders have argued that adopting English as an official language or adopting
English as the official language in certain workplace settings could ease labour market shortages and allow
countries to fill positions with qualified workers from other linguistic communities (Oltermann, 2023[47]).
However, this could result in the loss of a diversity that is strongly valued by the European Union as an
essential part of its democracy, resulting in its motto “Unity in Diversity”. Our data show that while English
is emerging as a lingua franca with the potential of aggregating the European labour market, national and
local languages retain a key role. The large majority of job ads are not published in English, but in the
national languages of each country’s labour market. In addition, local languages like Basque and Welsh
are demanded in hundreds of thousands of ads.
52. An important limitation of the study is that it is not possible, using information from job vacancies,
to assess the level of language proficiency requested. Although it is possible that some prospective
employers specify in the text of the vacancy a minimum or desired level of language proficiency,
information on the required level of proficiency is noisy and is present only in few vacancies. Therefore,
analyses presented simply considered if a vacancy required at least some knowledge of a specific
language either explicitly, because it mentioned a linguistic requirement in the vacancy text, or implicitly,
because the vacancy was written in that language (even though it may be non-official language in the
country in which the position was expected to be filled). This is an important limitation because the literature
indicates that the earning differentials associated with language skills are considerable but vary according
to a worker’s level of proficiency (Di Paolo and Tansel, 2015[48]; Liwiński, 2019[25]; Stöhr, 2015[49])
53. A second limitation of the study is that no estimates were developed to assess the wage returns
associated with language skills. This was due to the fact that except for the United Kingdom in which a
large number of vacancies contains information on the wages offered by prospective employers, in most

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  43

other countries information on prospective wage is missing from a large majority of job vacancies (because
it was not mentioned in the vacancy text). The literature on the wage returns associated with language
proficiency indicate that these are large but depend on the specific language workers are able to operate
in and the country in which they work. For example, in a large review on the wage returns to language
skills in European countries suggest that earning differentials associated with language skills range
between 6 and 50% (Ginsburgh and Prieto-Rodriguez, 2011[22]).
54. It is possible that in the future, artificial intelligence-(AI) powered machine translation tools will ease
the possibility for individuals with limited linguistic proficiency to live and work in other countries in Europe
breaking down some linguistic barriers in the labour market, especially in some occupations. Machine
translation technologies today can in fact translate some texts with a high level of accuracy. At the same
time, the quality of translations produced by AI remains variable and depends on the algorithms being used
by the language models, the quantity, quality, and variety of the translations used to train the machine
learning algorithms that are at the basis of AI machine translation tools as well as the complexity of the
text that is translated. Furthermore, although AI machine translation tools produce translations of
reasonable quality for written text, challenges related to real time synchronous interpretation with the added
challenges related to speech recognition put AI realistically in the future (Borgonovi, Hervé and Seitz,
2023[7]).
55. Technological advances shift the boundaries of what digital mediators can do. By doing so, they
change the opportunity cost for individuals to invest in language skills but also, potentially, of what specific
aspects of language they should acquire. Should AI make significant progress in facilitating inter-linguistic
understanding, it is possible that individuals put more and not less efforts in language learning, in order to
be able to understand and use the cultural significance of languages in their work and social exchanges.
Moreover, as the use of English becomes widespread among non-native users with a diverse range of
proficiency and ability to use the language productively and/or receptively in its written and oral form, a
unique skill will be to be able to make the most of communication with such diverse group of users.

Unclassified
44  DELSA/ELSA/WD/SEM(2023)12

Annex A. Robustness analysis on the regional

distribution of the demand for English

Figure A A.1. The unweighted regional distribution of online job vacancies requiring at least some
knowledge of English in 2021, by NUTS-3 region
Proportion by NUTS-3 regions, 2021

Note: The tables underlying all the charts in this paper are available in the GitHub repository created by Marconi and Vergolini (2022[37]).
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  45

Annex B. Occupational differences in the

demand for language skills

Table A B.1. The demand for French skills among managers, professionals, technicians, and
associate professionals
Percentage of online vacancies explicitly or implicitly requiring at least some knowledge of French among managers,
professionals, technicians and associate professionals in 2021, by country

Percentage of online job vacancies requiring French skills

Managers Professionals Technicians National average
Austria 0.086 0.036 0.265 0.118
Bulgaria 0.020 0.433 0.689 0.295
Croatia 0.388 0.165 50.007 15.096
Cyprus 0.120 0.142 0.349 0.159
Czech Republic 0.005 0.677 0.670 0.638
Denmark 0.013 0.028 0.022 0.019
Estonia 0.010 0.022 0.012 0.010
EU 2.066 2.856 4.824 3.347
Finland 0.047 0.041 0.056 0.044
Germany 1.183 0.608 0.771 0.647
Greece 0.148 0.281 0.690 0.615
Hungary 0.065 0.304 0.167 0.292
Ireland 1.550 1.674 0.691 1.296
Italy 2.918 2.366 2.032 2.132
Latvia 0.000 0.006 0.000 0.003
Lithuania 0.004 0.004 0.003 0.003
Luxembourg 33.205 51.719 57.469 53.895
Malta 0.081 0.056 0.000 0.029
Netherlands 2.322 1.498 0.757 0.916
OECD 2.432 3.344 3.350 3.269
Poland 0.682 1.458 0.930 0.844
Portugal 4.677 5.710 2.884 4.091
Romania 1.470 2.313 0.547 0.978
Slovak Republic 0.235 0.256 0.407 0.356
Slovenia 0.033 0.018 0.000 0.009
Spain 2.172 1.279 1.070 1.044
Sweden 0.213 0.305 0.123 0.147
United Kingdom 1.497 1.942 1.325 1.521

Note: Countries are sorted alphabetically. Percentage of vacancies either specifying French as a required skill in the vacancy text or vacancies
that were written in French. Data from France and Belgium were excluded. “National average” refers to the average in a specific country across
all occupational groups.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
46  DELSA/ELSA/WD/SEM(2023)12

Table A B.2. The demand for French skills among other occupations
Percentage of online vacancies explicitly or implicitly requiring at least some knowledge of French among clerical
support workers, service and sales workers, skilled agricultural, forestry and fishery workers, craft and related trades
workers, plant and machine operators and assemblers, and elementary occupations, by country

Percentage of online job vacancies requiring French skills

Clerical Service Agriculture Craft Plant Elementary National average
Austria 0.075 0.137 0.000 0.149 0.099 0.071 0.118
Bulgaria 0.083 0.075 0.000 0.109 0.019 0.122 0.295
Croatia 0.366 3.069 0.000 0.113 0.103 0.106 15.096
Cyprus 0.000 0.125 0.000 0.000 0.000 0.186 0.159
Czech Republic 0.021 0.016 0.000 0.010 2.821 0.009 0.638
Denmark 0.020 0.020 0.000 0.006 0.007 0.024 0.019
Estonia 0.000 0.007 0.000 0.000 0.000 0.000 0.010
EU 2.813 3.556 2.772 3.531 2.247 3.534 3.403
Finland 0.132 0.053 0.000 0.006 0.010 0.041 0.044
Germany 0.819 0.305 0.035 0.109 0.740 0.722 0.647
Greece 0.661 0.849 0.575 1.251 0.776 1.939 0.615
Hungary 1.062 0.302 0.000 0.100 0.169 0.261 0.292
Ireland 1.529 1.010 0.076 1.577 0.050 0.525 1.296
Italy 5.269 1.796 11.950 0.685 0.487 0.888 2.132
Latvia 0.000 0.000 0.000 0.000 0.012 0.000 0.003
Lithuania 0.000 0.006 0.000 0.000 0.000 0.000 0.003
Luxembourg 44.864 74.498 55.769 80.183 48.818 79.348 53.895
Malta 0.000 0.000 0.000 0.000 0.000 0.000 0.029
Netherlands 0.910 0.524 0.089 0.224 0.171 0.061 0.916
OECD 3.320 4.071 3.300 4.189 2.663 4.160 3.269
Poland 1.113 0.332 0.000 0.278 0.001 0.003 0.844
Portugal 9.220 3.335 0.258 1.859 0.257 0.638 4.091
Romania 0.234 0.258 0.000 0.230 0.233 0.631 0.978
Slovak Republic 0.586 0.243 0.267 0.376 0.212 0.733 0.356
Slovenia 0.000 0.029 0.000 0.000 0.000 0.000 0.009
Spain 1.832 0.932 0.126 0.190 0.222 0.673 1.044
Sweden 0.025 0.014 0.000 0.004 0.004 0.008 0.147
United Kingdom 1.580 1.091 0.151 0.967 1.060 1.425 1.521

Note: Countries are sorted alphabetically. Percentage of vacancies either specifying French as a required skill in the vacancy text or vacancies
that were written in French. Data from France and Belgium were excluded. “National average” refers to the average in a specific country across
all occupational groups.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  47

Table A B.3. The demand for German skills among managers, professionals, technicians, and
associate professionals
Percentage of online vacancies explicitly or implicitly requiring at least some knowledge of German among
managers, professionals and technicians and associate professionals in 2021, by country

Percentage of online job vacancies requiring German skills

Professionals Technicians Professionals National average
Belgium 11.336 9.708 8.843 7.403
Bulgaria 9.112 6.835 6.508 5.671
Croatia 0.427 0.433 49.672 15.175
Cyprus 0.401 0.632 1.148 0.884
Czech Republic 5.516 4.138 6.292 4.222
Denmark 2.312 0.073 0.059 0.305
Estonia 0.381 0.774 1.462 0.600
EU 3.801 3.163 5.164 3.305
Finland 0.118 0.125 0.204 0.210
France 2.365 1.873 1.539 1.238
Greece 1.119 1.918 2.516 2.149
Hungary 10.964 8.029 7.617 6.526
Ireland 1.980 1.397 1.068 1.387
Italy 2.264 1.578 0.997 1.217
Latvia 0.571 1.278 1.391 0.970
Lithuania 2.440 0.052 1.376 1.002
Luxembourg 21.031 17.455 22.519 17.689
Malta 0.243 0.225 0.281 0.321
Netherlands 3.380 2.115 1.749 1.700
OECD 3.977 3.250 3.390 2.854
Poland 4.481 5.529 4.523 4.306
Portugal 4.781 4.300 3.993 4.223
Romania 3.184 4.371 1.654 2.078
Slovak Republic 0.336 2.916 1.296 1.084
Slovenia 2.222 1.277 0.928 0.728
Spain 2.631 1.394 1.110 1.033
Sweden 1.431 0.641 0.356 0.516
United Kingdom 1.861 1.674 1.355 1.429

Note: Countries are sorted alphabetically. Percentage of vacancies either specifying German as a required skill in the vacancy text or vacancies
that were written in German. Data from Austria and Germany were excluded. “National average” refers to the average in a specific country
across all occupational groups.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
48  DELSA/ELSA/WD/SEM(2023)12

Table A B.4. The demand for German skills among other occupations
Percentage of online vacancies explicitly or implicitly requiring at least some knowledge of German among clerical
support workers, service and sales workers, skilled agricultural, forestry and fishery workers, craft and related trades
workers, plant and machine operators and assemblers, and elementary occupations, by country

Percentage of online job vacancies requiring German skills

Clerical Service Agriculture Craft Plant Elementary National average
Belgium 5.911 5.651 4.823 3.480 5.889 2.738 7.403
Bulgaria 6.894 1.873 0.746 2.603 8.995 3.483 5.671
Croatia 0.258 3.983 0.264 0.058 0.090 0.106 15.175
Cyprus 0.493 0.938 15.789 2.130 1.131 1.671 0.884
Czech Republic 5.570 1.650 0.000 1.337 6.851 1.227 4.222
Denmark 0.153 0.024 0.043 0.008 0.073 0.213 0.305
Estonia 0.999 0.204 0.606 0.528 0.295 0.279 0.600
EU 2.926 2.310 3.700 1.329 2.066 1.376 3.305
Finland 0.132 0.084 0.000 0.448 0.414 0.534 0.210
France 1.789 1.612 0.439 0.162 0.398 0.064 1.238
Greece 1.611 2.081 2.874 3.797 3.692 2.758 2.149
Hungary 7.620 6.255 3.835 1.195 2.699 1.369 6.526
Ireland 2.359 1.091 0.534 1.767 0.041 0.215 1.387
Italy 3.046 1.020 5.346 0.126 0.316 0.240 1.217
Latvia 1.085 1.600 0.254 0.222 0.160 0.435 0.970
Lithuania 0.560 0.753 0.543 0.086 3.526 0.260 1.002
Luxembourg 13.403 16.362 44.231 7.695 8.473 13.473 17.689
Malta 0.198 0.921 0.000 0.353 0.438 0.602 0.321
Netherlands 1.413 1.526 2.517 0.957 0.884 1.007 1.700
OECD 3.171 2.417 3.483 1.337 1.872 1.350 2.854
Poland 5.282 3.430 0.830 2.893 3.104 0.013 4.306
Portugal 10.967 4.685 2.696 1.626 0.799 1.124 4.223
Romania 0.189 0.383 3.514 0.668 2.648 0.827 2.078
Slovak Republic 0.719 0.221 0.801 0.486 0.225 0.549 1.084
Slovenia 0.409 0.251 1.600 0.433 0.330 0.716 0.728
Spain 1.412 0.679 0.151 0.146 0.126 0.467 1.033
Sweden 0.686 0.462 0.065 0.016 0.055 0.029 0.516
United Kingdom 1.458 1.109 0.961 0.678 0.963 0.645 1.429

Note: Countries are sorted alphabetically. Percentage of vacancies either specifying German as a required skill in the vacancy text or vacancies
that were written in German. Data from Austria and Germany were excluded. “National average” refers to the average in a specific country
across all occupational groups.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  49

Table A B.5. The demand for Spanish skills among managers, professionals, technicians and
associate professionals
Percentage of online vacancies explicitly or implicitly requiring at least some knowledge of Spanish among
managers, professionals and technicians and associate professionals in 2021, by country

Percentage of online job vacancies requiring Spanish skills

Managers Professionals Technicians National average
Austria 4.885 0.031 0.003 0.015
Belgium 1.943 0.012 0.016 0.010
Bulgaria 5.472 0.044 0.020 0.025
Croatia 0.000 0.056 0.065 0.150
Cyprus 0.000 0.001 0.000 0.001
Czech Republic 0.839 0.017 0.013 0.008
Denmark 0.000 0.000 0.000 0.000
Estonia 0.010 0.000 0.000 0.000
EU 2.687 0.030 0.025 0.027
Finland 0.008 0.000 0.000 0.000
France 2.178 0.016 0.016 0.012
Germany 6.802 0.030 0.017 0.021
Greece 0.082 0.002 0.003 0.005
Hungary 1.417 0.030 0.010 0.015
Ireland 2.145 0.019 0.018 0.018
Italy 1.667 0.010 0.004 0.006
Latvia 0.000 0.000 0.000 0.000
Lithuania 0.064 0.002 0.002 0.002
Luxembourg 4.628 0.121 0.095 0.078
Malta 0.081 0.000 0.001 0.001
Netherlands 2.052 0.010 0.009 0.008
OECD 2.836 0.030 0.025 0.024
Poland 3.015 0.039 0.024 0.020
Portugal 28.819 0.298 0.306 0.289
Romania 3.015 0.035 0.034 0.022
Slovak Republic 0.000 0.015 0.000 0.003
Slovenia 0.000 0.000 0.000 0.000
Sweden 0.736 0.004 0.000 0.002
United Kingdom 1.097 0.010 0.006 0.008
Austria 4.885 0.031 0.003 0.015
Belgium 1.943 0.012 0.016 0.010

Note: Countries are sorted alphabetically. Percentage of vacancies either specifying Spanish as a required skill in the vacancy text or vacancies
that were written in Spanish. Data from Spain were excluded. “National average” refers to the average in a specific country across all
occupational groups.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
50  DELSA/ELSA/WD/SEM(2023)12

Table A B.6. The demand for Spanish skills among other occupations
Percentage of online vacancies explicitly or implicitly requiring at least some knowledge of Spanish among clerical
support workers, service and sales workers, skilled agricultural, forestry and fishery workers, craft and related trades
workers, plant and machine operators and assemblers, and elementary occupations, by country

Percentage of online job vacancies requiring Spanish skills

Clerical Service Agriculture Craft Plant Elementary National average
Austria 1.223 0.009 0.000 0.013 1.130 0.005 1.499
Belgium 0.930 1.155 0.000 0.011 0.812 0.015 1.047
Bulgaria 4.006 0.729 0.015 0.104 0.095 0.037 2.460
Croatia 15.983 47.419 0.264 15.617 15.068 5.088 15.012
Cyprus 0.000 0.000 0.000 0.290 0.323 0.371 0.064
Czech Republic 0.003 0.004 0.000 0.005 1.290 0.001 0.834
Denmark 0.015 0.003 0.000 0.006 0.000 0.005 0.005
Estonia 0.000 0.013 0.000 0.020 0.015 0.000 0.020
EU 2.892 3.480 0.492 1.841 2.341 0.416 2.736
Finland 0.029 0.031 0.000 0.019 0.000 0.010 0.015
France 1.866 1.513 0.094 0.287 0.787 0.068 1.210
Germany 1.771 1.168 0.012 0.191 1.383 0.004 2.095
Greece 2.768 0.157 0.575 0.931 0.752 0.205 0.512
Hungary 1.245 0.844 0.197 0.125 0.122 0.079 1.466
Ireland 3.291 1.342 0.000 0.061 1.018 0.045 1.789
Italy 0.963 0.353 4.403 0.103 0.262 0.036 0.578
Latvia 0.013 0.000 0.000 0.000 0.000 0.000 0.001
Lithuania 0.268 0.388 0.000 0.275 0.398 0.371 0.230
Luxembourg 0.000 0.000 0.000 0.000 0.000 0.000 7.811
Malta 0.000 0.132 0.000 0.000 0.146 0.000 0.057
Netherlands 1.006 0.630 0.000 0.020 0.137 0.024 0.754
OECD 2.533 1.897 0.384 1.320 2.016 0.197 2.369
Poland 1.598 0.445 0.000 0.002 0.419 0.007 2.024
Portugal 37.143 33.067 3.028 26.870 35.552 3.426 28.896
Romania 0.615 1.051 4.054 2.863 1.109 0.980 2.204
Slovak Republic 0.009 0.012 0.134 0.029 0.000 0.020 0.330
Slovenia 0.030 0.000 0.000 0.000 0.000 0.000 0.004
Sweden 0.410 0.014 0.011 0.014 0.034 0.012 0.223
United Kingdom 1.138 0.583 0.000 0.059 0.247 0.011 0.762

Note: Countries are sorted alphabetically. Percentage of vacancies either specifying Spanish as a required skill in the vacancy text or vacancies
that were written in Spanish. Data from Spain were excluded. “National average” refers to the average in a specific country across all
occupational groups.
Source: Web Intelligence Hub’s Online Job Advertisement (OJA) release “2021q4” database.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  51

References

Abbott, D. (2014), Applied Predictive Analytics: Principles and Techniques for the Professional [44]
Data Analyst, Wiley.

Antonietti, R. and M. Loi (2014), “Firm Internationalization and Demand for Foreign Language [28]
Skills: A Firm-level Investigation.”, Economia Politica, Vol. 2, pp. 173-194.

Araújo, L. et al. (2015), Languages and employability, Joint Research Centre, [20]
http://dx.doi.org/10.2788/860807.

Armstrong, A. (2015), “Equilibria and efficiency in bilingual labour markets”, Journal of Economic [26]
Behavior &amp; Organization, Vol. 112, pp. 204-220,
https://doi.org/10.1016/j.jebo.2015.01.011.

Ascheri, A. et al. (2022), “Competition in urban hiring markets: evidence from online job [30]
advertisements.”, Eurostat Statistical Working Papers, https://doi.org/10.2785/667004.

Bleakley, H. and A. Chin (2004), “Language Skills and Earnings: Evidence from Childhood [13]
Immigrants*”, Review of Economics and Statistics, Vol. 86/2, pp. 481-496,
https://doi.org/10.1162/003465304323031067.

Borgonovi, F. and A. Ferrara (2020), “Academic achievement and sense of belonging among [10]
non-native-speaking immigrant students: The role of linguistic distance”, Learning and
Individual Differences, Vol. 81, p. 101911, https://doi.org/10.1016/j.lindif.2020.101911.

Borgonovi, F., J. Hervé and H. Seitz (2023), “Not lost in translation: The implications of machine [7]
translation technologies for language professionals and for broader society”, OECD Social,
Employment and Migration Working Papers, No. 291, OECD Publishing, Paris,
https://doi.org/10.1787/e1d1d170-en.

Cammeraat, E. and M. Squicciarini (2021), “Burning Glass Technologies’ data use in policy- [34]
relevant analysis: An occupation-level assessment”, OECD Science, Technology and Industry
Working Papers, No. 2021/05, OECD Publishing, Paris, https://doi.org/10.1787/cd75c3e7-en.

Cedefop (2022), Skills Online Vacancy Analysis Tool for Europe, [33]
https://www.cedefop.europa.eu/en/tools/skills-online-vacancies/skills/occupations.

Cedefop (2019), Online Job Vacancies and Skills Analysis – A Cedefop Pan-European [31]
Approach, Publications Office, Luxembourg, http://data.europa.eu/doi/10.2801/097022.

Unclassified
52  DELSA/ELSA/WD/SEM(2023)12

Chen, S. (2022), Employers’ language requirements towards international students in the pre- [36]
recruitment stage: The case of Aalto University in Finland, Kauppakorkeakoulu School of
Business, https://aaltodoc.aalto.fi/handle/123456789/115138?show=full.

Chiswick, B. and P. Miller (2003), “The complementarity of language and other human capital: [8]
immigrant earnings in Canada”, Economics of Education Review, Vol. 22/5, pp. 469-480,
https://doi.org/10.1016/s0272-7757(03)00037-2.

Chiswick, B. and P. Miller (1995), “The Endogeneity between Language and Earnings: [9]
International Analyses”, Journal of Labor Economics, Vol. 13/2, pp. 246-288,
https://doi.org/10.1086/298374.

Commission, E. (n.d.), https://cros-legacy.ec.europa.eu/content/trusted-smart-statistics- [32]

%E2%80%93-web-intelligence-hub_en.

Council of the European Union (2014), Council conclusions of 20 May 2014 on multilingualism [16]
and the development of language competences, https://op.europa.eu/en/publication-detail/-
/publication/18fb5e89-f38a-11e3-831f-01aa75ed71a1/language-en.

Della Chiesa, B., J. Scott and C. Hinton (eds.) (2012), Languages in a Global World: Learning for [1]
Better Cultural Understanding, Educational Research and Innovation, OECD Publishing,
Paris, https://doi.org/10.1787/9789264123557-en.

Di Paolo, A. and A. Tansel (2015), “Returns to Foreign Language Skills in a Developing Country: [48]
The Case of Turkey”, The Journal of Development Studies, Vol. 51/4, pp. 407-421,
https://doi.org/10.1080/00220388.2015.1019482.

Dustmann, C. and F. Fabbri (2003), “Language Proficiency and Labour Market Performance of [14]
Immigrants in the UK”, The Economic Journal, Vol. 113/489, pp. 695-717,
https://doi.org/10.1111/1468-0297.t01-1-00151.

Dustmann, C. and A. Soest (2001), “Language Fluency and Earnings: Estimation with [15]
Misclassified Language Indicators”, Review of Economics and Statistics, Vol. 83/4, pp. 663-
674, https://doi.org/10.1162/003465301753237740.

Eberhard, D. et al. (2023), Ethnologue: Languages of the World, [39]

https://www.ethnologue.com/guides/ethnologue200.

Eurostat (2017), Foreign language learning, [45]

https://ec.europa.eu/eurostat/documents/2995521/7879483/3-23022017-AP-
EN.pdf/80715559-72ba-4c19-b341-7ddb42dd61a6.

Eurostat (n.d.), Foreign language skills statistics, https://ec.europa.eu/eurostat/statistics- [46]

explained/index.php?title=Foreign_language_skills_statistics#Analysis_of_those_knowing_on
e_or_more_foreign_languages.

Fabo, B., M. Beblavý and K. Lenaerts (2017), “The importance of foreign language skills in the [29]
labour markets of Central and Eastern Europe: assessment based on data from online job
portals”, Empirica, Vol. 44/3, pp. 487-508, https://doi.org/10.1007/s10663-017-9374-6.

Gazzola, M. and D. Mazzacani (2019), “Foreign language skills and employment status of [21]
European natives: evidence from Germany, Italy and Spain”, Empirica, Vol. 46/4, pp. 713-
740, https://doi.org/10.1007/s10663-019-09460-7.

Unclassified
 DELSA/ELSA/WD/SEM(2023)12  53

Ginsburgh, V. and J. Prieto-Rodriguez (2011), “Returns to Foreign Languages of Native Workers [22]
in the European Union”, ILR Review, Vol. 64/3, pp. 599-618,
https://doi.org/10.1177/001979391106400309.

Grin, F., C. Sfreddo and F. Vaillancourt (2011), The Economics of the Multilingual Workplace, [17]
Routledge, https://doi.org/10.4324/9780203852675.

Gudykunst, W. (2003), Cross-Cultural and Intercultural Communication, Sage Publishing. [27]

Guven, C. and A. Islam (2015), “Age at Migration, Language Proficiency, and Socioeconomic [12]
Outcomes: Evidence From Australia”, Demography, Vol. 52/2, pp. 513-542,
https://doi.org/10.1007/s13524-015-0373-6.

Hahm, S. and M. Gazzola (2022), “The Value of Foreign Language Skills in the German Labor [23]
Market”, Labour Economics, Vol. 76, p. 102150,
https://doi.org/10.1016/j.labeco.2022.102150.

Hummels, D. (2007), “Transportation Costs and International Trade in the Second Era of [3]
Globalization”, Journal of Economic Perspectives, Vol. 21/3, pp. 131-154,
https://doi.org/10.1257/jep.21.3.131.

ILO (2012), International Standard Classification of Occupations, [38]

https://www.ilo.org/wcmsp5/groups/public/@dgreports/@dcomm/@publ/documents/publicatio
n/wcms_172572.pdf.

Isphording, I. (2015), “Language and Labor Market Success”, in International Encyclopedia of [24]
the Social &amp; Behavioral Sciences, Elsevier, https://doi.org/10.1016/b978-0-08-097086-
8.94033-9.

Isphording, I. and S. Otten (2014), “Linguistic barriers in the destination language acquisition of [5]
immigrants”, Journal of Economic Behavior &amp; Organization, Vol. 105, pp. 30-50,
https://doi.org/10.1016/j.jebo.2014.03.027.

Kiškytė, M. (2023), “Demands of the Employers for Foreign Language Skills in Job [35]
Advertisements in Three Baltic Countries”, Socialiniai tyrimai, Vol. 46/1, pp. 24-52,
https://doi.org/10.15388/soctyr.2022.46.1.2.

Liwiński, J. (2019), “The wage premium from foreign language skills”, Empirica, Vol. 46/4, [25]
pp. 691-711, https://doi.org/10.1007/s10663-019-09459-0.

Marconi, G. and L. Vergolini (2022), “The Demand for Language Skills in the European Labour [37]
Market: Evidence From Online Job Ads”, SSRN Electronic Journal,
https://doi.org/10.2139/ssrn.4152463.

Martel, P. and J. Maurais (eds.) (1994), Langues et sociétés en contact, DE GRUYTER, [18]
https://doi.org/10.1515/9783110932935.

Melitz, J. (2008), “Language and foreign trade”, European Economic Review, Vol. 52/4, pp. 667- [4]
699, https://doi.org/10.1016/j.euroecorev.2007.05.002.

Montgomery, S. (2013), Does Science Need a Global Language? English and the Future of [40]
Research, The University of Chicago Press.

Unclassified
54  DELSA/ELSA/WD/SEM(2023)12

OECD (2018), Skills on the Move: Migrants in the Survey of Adult Skills, OECD Skills Studies, [11]
OECD Publishing, Paris, https://doi.org/10.1787/9789264307353-en.

Oltermann, P. (2023), “German call for English to be second official language amid labour [47]
shortage”, The Guardian, https://www.theguardian.com/world/2023/feb/10/germany-labour-
shortage-english-second-official-language.

Ovchinnikova, E., C. Mol and E. Jones (2022), “The role of language proximity in shaping [6]
international student mobility flows”, Globalisation, Societies and Education, pp. 1-12,
https://doi.org/10.1080/14767724.2022.2070132.

Plewa, P. and M. Stermšek (2017), Labour Migration from China to Europe: Scope and [42]
Potential., http://www.ilo.int/wcmsp5/groups/public/---asia/---ro-bangkok/---ilo-
beijing/documents/publicatio.

Poldrack, R., G. Huckins and G. Varoquaux (2020), “Establishment of Best Practices for [43]
Evidence for Prediction”, JAMA Psychiatry, Vol. 77/5, p. 534,
https://doi.org/10.1001/jamapsychiatry.2019.3671.

Spence, P. and G. Liu (2013), “Engineering English and the high-tech industry: A case study of [41]
an English needs analysis of process integration engineers at a semiconductor manufacturing
company in Taiwan”, English for Specific Purposes, Vol. 32/2, pp. 97-109,
https://doi.org/10.1016/j.esp.2012.11.003.

Stöhr, T. (2015), “The returns to occupational foreign language use: Evidence from Germany”, [49]
Labour Economics, Vol. 32, pp. 86-98, https://doi.org/10.1016/j.labeco.2015.01.004.

Tabouret-Keller, A. (2017), “Language and Identity”, in The Handbook of Sociolinguistics, [2]

Blackwell Publishing Ltd, Oxford, UK, https://doi.org/10.1002/9781405166256.ch19.

Vaillancourt, F. (1996), “Language and socioeconomic status in Quebec: measurement, findings, [19]
determinants, and policy costs”, International Journal of the Sociology of Language,
Vol. 121/1, https://doi.org/10.1515/ijsl.1996.121.69.

Unclassified