Substitution risk and continuing training behaviour in the digital shift
What are the gender-specific differences?
Professional knowledge is ageing ever more quickly in the course of the digital shift, and continuing vocational training measures are becoming increasingly significant in terms of preparing the labour force for new and/or altered task requirements. The extent to which occupations typically exercised by men are affected differently to occupations in which women are in the majority remains unclear at present. This article conducts an investigation at the level of occupational groups in order to discover whether there is any evidence of a gender-specific risk of being affected by digital substitution processes and whether the previous continuing training behaviour displayed by the labour force has been adapted to digital structural change.
Can continuing training prevent substitution risks?
As a result of the digital shift, task requirements are changing or are in some cases being entirely replaced by technical innovations (cf. Dengler/Matthes 2018). The degree of substitution particularly depends on the proportion of creative, interactive, routine and/or autonomous task elements in an occupation (cf. Bonin/Gregory/Zierahn 2015). Structural shifts in work and services processes have previously predominantly led to the replacement of simple cognitive routine tasks (cf. Autor/Levy/Murnane 2003). New tasks mostly emerge to complement technical innovations, meaning that individual creativity and human interaction, innovation and decision-making capacities are gaining in significance amidst uncertainty and in complex contexts (cf. Eichhorst/Buhlmann 2018). Helmrich et al. (2016) were also able to demonstrate empirically that digitalisation is fostering accelerated job switching and structural change on the German labour market. The present article assumes that structural change causes the employees either to take on new tasks which tend to be more demanding or in certain circumstances forces them to retrain for a different task. Occupational groups which are under particularly severe threat from substitution ought therefore to display a greater scope of continuing training in order to react to or be prepared for the task changes. Nevertheless, occupation-centred empirical research has suggested that substitution effects take place at different times and at various paces in the respective occupations (cf. KRUPPE et al. 2019).
Gender as a “blind spot” of substitution research
Because of the markedly gender-specific structuring of the German labour market (cf. Hausmann/Kleinert 2014), the question which emerges is the extent to which “typical” male and female occupations respectively are threatened by substitution. Gender differences are frequently not taken into account in the current discourse surrounding digitalisation, and research activities tend to focus on male occupations (e.g. the craft trades/technical occupations) (cf. Kutzner 2017). This inevitably leads to a one-sided and “male” interpretation of the effects of digitalisation. However, initial empirical research indicates that women tend to evaluate the effects of digitalisation more negatively than men and that they express different requirements and concerns (cf. Seegers 2018). This necessitates a gender-sensitive analysis of the effects of digitalisation.
Research questions and data base
The thesis propounded here is that task requirements in male, female and mixed occupations are changing differently in the course of digitalisation or that they are threatened in different ways by substitution. This is particularly significant when appropriate continuing vocational training measures need to be designed in order to adapt to the new task requirements. For this reason, the aim of the investigation is to identify occupations under particular threat from substitution by analysing the relevant task proportions. These are then analysed by occupational group and contrasted with the respective continuing training behaviour exhibited by workers. Particular focus is placed on the gender-segregated strategy used for the analysis, and this will be explained in the following section. The article is based on the following research questions. Are male, female and mixed occupations under differing degrees of threat from substitution? Which task proportions foster the substitution risk? And to which extent does each occupational group take recourse to continuing vocational training? The database used is the Adult Survey carried out by the German National Educational Panel Study (NEPS, start cohort 6, 2015/2016, 2016/2017 waves, cf. Blossfeld/Roßbach/maurice 2011). Microcensus data from the period 2015–2017 relating to the gender proportion per occupational group has additionally been deployed (cf. FDZ 2016) in order to form gender-typing male, female and mixed occupational categories (cf. Information Box).
The National Educational Panel Study (NEPS) offers longitudinal data on aspects such as competence developments and education and training processes in formal, non-formal and informal contexts across the entire lifespan. Data is collected by means of a representative sample carried out by the Einwohnermeldeamt [Residents’ Registration Office].
The microcensus survey was used to form gender-typical occupational category groups on the basis of the average proportion of women per occupational group in the years 2015–2017.
Proportion of women ≤ 30 % = male occupational group
Proportion of women ≥ 70 % = female occupational groups
Proportion of women > 30 % & < 70 % = mixed occupational groups
Occupational groups and task requirements
The coding of the “2010 Classification of occupations” (KldB 2010)1 was used to link the gender-typing categories formed from the microcensus with the NEPS data at the level of occupational groups.2 An occupational group and a gender-typing occupational category is thus allocated to each worker. In the interests of ease of legibility, the occupational groups are referred to below simply as occupations. The analysis covered a total of 8,908 workers aged between 18 and 75 from 67 occupational groups and from two waves of the survey (2015/16, 2016/17). Occupational groups with a sample size of less than 30 were excluded because of their limited significance.
The NEPS data encompasses measurements relating to task proportions (varied, autonomous, interactive) and to scope of continuing training (hours of continuing training per person in the period 2015–2017). An exploratory factor analysis was carried out to calculate an index from the individual task proportions. Each person receives a value for the respective task. This may be positive or negative. Negative factor values indicate a below-average task proportion. Positive factors show an above-average proportion, and values approaching zero designate an average proportion. The factor values were then aggregated to form a common index. The lower this common index, the higher the substitution risk of the occupations will be. This overall index is used to identify and analyse the five occupations per category with the highest substitution risk from the original total of 67 occupational groups. This means that the final analysis is informed by a total employees of 1,500 persons.
Substitution risk, task proportions and further training of employees
The Table presents the task proportions and the scope of continuing training for employees in the five occupations with the highest substitution risk, differentiated by male, female and mixed occupations.
The male occupations analysed generally exhibit the highest substitution risk in overall terms (-8.38). They also display the lowest scope of continuing training per employed person, 60.6 hours on average. The female occupations show the second highest substitution risk overall (-7-17). Average scope of participation in continuing training in this case is 117.2 hours per person, almost double the corresponding figure for those working in male occupations. The lowest substitution risk (-5.00) can be demonstrated for the mixed occupations listed, in which workers complete an average of 95.6 hours of continuing vocational training measures per employee.
Employees in female occupations stand out by dint of a low proportion of varied tasks (-3.83) but also due to a comparatively high proportion of interactive tasks (-1.43). Cleaning and housekeeping occupations are at most risk of substitution because of their relatively low proportion of varied (-0.99/-1.21) and interactive tasks (-0.76/-0.99). Cleaning staff display the lowest scope of continuing training per employee, an average of 79 hours. Those working as doctors’ receptionists and assistants have completed an average of 175 hours of continuing training. This is the highest figure recorded for any of the cross-gender occupational groups analysed. It should also be noted that two of the five female occupations are allocated to the main occupational group of retail. They exhibit a particular substitution risk because of their relatively low proportion of autonomous tasks and lack of variety.
All in all, the mixed occupations analysed revealed a comparatively high proportion of varied tasks (-1.09). However, the proportion of interactive tasks emerges as a negative element (-2.73). Warehousing and postal occupations are most at threat from substitution (-1.86), followed by occupations in gastronomy. Occupations in cooking and food preparation show the lowest scope of continuing training, an average of 53 hours per worker. Chemical occupations and occupations in technical drawing display the lowest substitution risk because of their comparatively high proportion of varied tasks (0.15 and 0.25 respectively). Both these occupations show a moderate amount of continuing training per worker of 114 hours and 124 hours respectively.
In overall terms, the five male occupations analysed are conspicuous by their low proportion of interactive tasks (ranging from 0.66 to 0.96, -4.06 in total) and by a small scope of continuing training per worker (38 to 75 hours and an average of 60.6 hours). Employees in occupations in the production of foodstuffs, confectionery and tobacco products have the lowest amount of continuing training of any of the cross-gender occupational groups analysed, an average of 38 hours. Drivers of vehicles in road traffic are most at risk from substitution (-2.14). The low proportion of autonomy in such occupations (-0.81) is a particular factor in this regard. Consideration also needs to be accorded to the fact that four of the five male occupations are from the occupational area of extraction of raw materials, production and manufacturing. Three of the occupational groups have their origin in the main occupational group of occupations in metal making, metal working and metal construction. From a horizontal perspective, metal production and manufacturing occupations are at particular threat from substitution.
Male occupations most at threat from substitution in overall terms
In conclusion, a gender-specific substitution risk to the detriment of male-dominated occupations can be identified. Although female-dominated occupations are also at threat, the risk is not as extensive and broad as in male occupations. In total, the male occupations analysed exhibit both a higher substitution risk and a smaller scope of continuing training per worker. One aspect that should be especially emphasised is that the average scope of continuing training of those employed in female occupations is almost twice as high on average as workers in male occupations.
In the male and female occupations considered, the substitution risk tends to be concentrated in individual occupational areas. In the mixed occupations, however, this threat is thinly spread across various occupational groups. The male-dominated metal working occupations, the mixed warehousing/postal occupations and the predominantly female retail occupations are exposed to the largest and most powerful threat from substitution because of the number of employees within the occupational groups and their high proportion of tasks that can be automated. According to current figures from the Statistisches Bundesamt [Federal Statistical Office], these three groups account for a total of 15.39 per cent of all employees subject to mandatory social insurance contributions. This highlights once more the dimension of the impending structural change.
The low total proportion of varied tasks in female occupations and the small amount of interactivity in male occupations are characteristic of these gender-typing occupational categories. If we focus on individual task elements and interpret activities with little variety as mainly being routine cognitive tasks, then tasks contained within typical female occupations are most likely to be replaced. However, because of the high amount of continuing training undertaken by workers in female occupations, the digital structural shift also provides a chance to upgrade “typical” female work both in terms of content and financially. This also represents an opportunity to break open and mix up the rigid gender segregation on some parts of the German labour market. At the same time, workers in male occupations should receive timely support from their employers and from policy makers in the form of continuing vocational training/retraining measures in order to arm them against the substitution effects which are already taking place, particularly in production and manufacturing occupations. As far as the specific organisation and implementation of continuing training measures are concerned, account should also be taken of gender-specific life circumstances (e.g. working time models, availability of childcare, geographic mobility) in order to create egalitarian opportunities for participation.
The cross-sectional design of the analysis means that it is only possible to arrive at a momentary conclusion. This illustrates the necessity for deeper longitudinal analyses.
An overall summary of the classifications is available at: https://statistik.arbeitsagentur.de/Statischer-Content/Grundlagen/Klassifikationen/Klassifikation-der-Berufe/KldB2010/Systematik-Verzeichnisse/Generische-Publikationen/Systematisches-Verzeichnis-Berufsbenennung.xls (retrieved: 29.11.2019).
The five hierarchical category levels of the KldB 2010 are occupational area (1-digit numbering), main occupational groups (2 digits), occupational groups (3 digits), occupational sub-groups (4 digits) and types of occupation (5 digits).
AUTOR, D.; LEVY, F.; MURNANE, R.: The Skill Content of Recent Technological Change: An Empirical Exploration. In: The Quarterly Journal of Economics 118 (2003) 4, pp. 1279-1333
BLOSSFELD, H.-P.; ROßBACH, H.-G.; Maurice, J.: Education as a lifelong process. The German National Educational Panel Study (NEPS). Wiesbaden 2011
BONIN, H.; GREGORY, T.; ZIERAHN, U.: Übertragung der Studie von Frey/Osborne (2013) auf Deutschland. Endbericht an das Bundesministerium für Arbeit und Soziales. Mannheim 2015
DENGLER, K.; MATTHES, B.: Substituierbarkeitspotenziale von Berufen: Wenige Berufsbilder halten mit der Digitalisierung Schritt (IAB-Kurzbericht 4/2018). Nürnberg 2018
EICHHORST, W.; BUHLMANN, F.: Wie gesellschaftlicher Fortschritt die Arbeitsorganisation prägt (IZA Standpunkte 91). Bonn 2018
FORSCHUNGSDATENZENTREN DER STATISTISCHEN ÄMTER DES BUNDES UND DER LÄNDER (FDZ): Mikrozensus. 2016. DOI: 10.21242/12211.2016.00.00.1.1.0
HAUSMANN, A.; KLEINERT, C.: Berufliche Segregation auf dem Arbeitsmarkt: Männer- und Frauendomänen kaum verändert (IAB-Kurzbericht 9/2014). Nürnberg 2014
KRUPPE, T. et al.: Digitalisierung, Herausforderungen für die Aus- und Weiterbildung in Deutschland (IAB-Stellungnahme 1/2019) Nürnberg 2019
KUTZNER, E.: Arbeit und Geschlecht. Die Geschlechterperspektive in der Auseinandersetzung mit Arbeit – aktuelle Fragen und Herausforderungen. Hans-Böckler-Stiftung. Düsseldorf 2017
SEEGERS, M.: Geschlechtsspezifische Rezeption von Digitalisierungsprozessen in Deutschland. Wie Geschlechterrollen die Wahrnehmung von Digitalisierungsrisiken und -chancen beeinflussen. Masterarbeit Universität Osnabrück 2018
Research Associate at the University of Osnabrück, Research Associate at BIBB
Translation from the German original (published in BWP 1/2020): Martin Kelsey, GlobalSprachTeam, Berlin