# Knowledge intensive business services: innovation and occupations

Ian Douglas Miles (Research Laboratory for Economics of Innovations, Institute for Statistical Studies and Economics of Knowledge, National Research University Higher School of Economics, Moscow, Russian Federation, and University of Manchester, Manchester, UK)
Veronika Belousova (Department of Educational Programmes/Institute for Statistical Studies and Economics of Knowledge, National Research University Higher School of Economics, Moscow, Russian Federation)
Nikolay Chichkanov (Unit for Intellectual Services Market Research/Institute for Statistical Studies and Economics of Knowledge, National Research University Higher School of Economics, Moscow, Russian Federation)

ISSN: 1463-6689

Publication date: 30 May 2019

## Abstract

### Purpose

The literature on knowledge-intensive business services (KIBS) shows them to be major innovators; this is confirmed with recent data, which the authors use to examine the various types of innovation that KIBS undertake. The implications for employment and work in highly innovative industries are important topics for analysis, not least because we are in a period where dramatic claims are being made as to the implications of new technologies for professional occupations. Thus, this paper aims to address major debates and conclusions concerning innovation patterns in KIBS and the evolving structures of professional and other work in these industries.

### Design/methodology/approach

This essay combines literature review with presentation and discussion of statistics that throw light on the patterns of innovation that characterise KIBS. The authors also consider data that concern trends in the organisation of work in these industries; while the focus is mainly on KIBS firms, they also pay some attention to KIBS-like work in other sectors. Even though KIBS are distinctive industries in modern economies, these analyses can be related to more general studies of, and forecasts about, changes in work organisation.

### Findings

The authors show that innovation patterns and employment structures vary substantially across different types of KIBS, with the distinction between technological, professional and creative KIBS proving to be useful for capturing these differences. The authors are also able to demonstrate important long- and medium-term trends in the structure and activities of the KIBS industries. In particular, data clearly demonstrate the increasing share of professional as against associate and clerical workers in most KIBS. Evidence also suggests that polarisation trends across the economy are mirrored, and in some cases amplified, in KIBS. The future prospects for employment in KIBS, and for professional work in particular, are seen to involve multiple factors, which together may bring about substantial change.

### Research limitations/implications

The study involves literature review and industry-level statistical analysis. Future work would benefit from firm-level analysis and validation and explication of results via consultation with practitioners and users of KIBS. Some puzzling variations across countries and sectors will need to be explored with national and sectoral experts.

### Practical implications

Research into KIBS activities, and their future, should make more use of the extensive statistics on employment and other structural features of the industries that have become available in recent years. KIBS firms and practitioners will need to take account of the forces for change that are liable to restructure their activities.

### Originality/value

The literature on KIBS has been concentrated on a rather narrow range of issues, while analysis of the current contributions and future development of the industries requires attention to a wider range of topics. This paper suggests how these topics may be investigated and their implications explored and presents results of enquiries along these lines.

## Citation

Miles, I.D., Belousova, V. and Chichkanov, N. (2019), "Knowledge intensive business services: innovation and occupations", Foresight, Vol. 21 No. 3, pp. 377-408. https://doi.org/10.1108/FS-11-2018-0091

## Publisher

:

Emerald Publishing Limited

## Introduction

Over a decade ago, this journal featured an essay on “knowledge intensive business services” (KIBS) (Miles, 2005 we refer to this as f2005 in the following discussion). This is the second in a series of three essays bringing the discussion up to date, reviewing major contributions to the study of KIBS, and considering emerging issues and their implications for the future of KIBS and related activities. The first of these three (Miles et al., 2018; we shall refer to this as f2018) discussed the nature of KIBS activities, the challenges of identifying these industries statistically and the extent to which studies over the past decade, and available empirical data, substantiate claims as to their importance and performance. The third essay will address the role of KIBS in innovation systems, especially considering how KIBS may (or may not) contribute to the transition to more sustainable paths of socioeconomic development.

This second paper addresses the controversial issue of the future of KIBS work and employment. Among studies claiming that new technologies are poised to transform work both radically and rapidly, we find several authors arguing that professional service activities will be facing major disruptive forces in the coming years. As in the other two essays, we can draw on reviews of the extensive literature on KIBS. However, there has been relatively little explicit discussion of several major issues that bear substantially on the current activity and future prospects for KIBS; we seek to offset this by presenting and analysing relevant data.

The essay begins by examining the oft-made point that KIBS are unusually innovative sectors of many major economies. The discussion focuses on innovation activities KIBS undertake to improve their own products, processes, organisational strategies or marketing activities. (The third of this set of essays considers KIBS’ role in fostering innovation across the whole economy.) Given current concerns about new technologies and employment prospects, we then turn to the topic of innovation and employment, paying attention to the structure of employment (e.g. its distribution across occupational categories, relevant to debates about deskilling and polarisation). We consider both forecasts and empirical research on these topics.

Readers are referred to f2018 for detailed explanation of the conceptualisation of KIBS, the statistical classifications of KIBS industries and the distinctions between P-KIBS, T-KIBS and C-KIBS (which are, respectively, specialised in professional, technological/scientific and creative/cultural knowledge).

## Innovation in knowledge-intensive business services

KIBS are widely held to be particularly innovative sectors of the economy: this theme of f2005 was reinforced in numerous subsequent studies and reviews. For example, f2018 noted that the systematic reviews of Scarso (2015), Braga and Marques (2016) and J-Figueiredo et al. (2017) all stressed this feature of the literature. Innovation should, in principle, result in KIBS providing clients with new and better services, and/or providing services at lower costs. It is also a major influence on the number of jobs, but the types of work undertaken and the combinations of occupations, in KIBS.

The high relative innovativeness of KIBS has been confirmed by various types of data, most often in Europe through the results of Community Innovation Survey (CIS) statistics covering most market sectors across EU countries. Two major studies using data from the CIS4 (conducted in 2004, covering the EU27) demonstrated this. Gotsch et al. (2011) stated that the share of innovative firms in the EU KIBS sector is higher than in manufacturing. The average share of innovative firms in KIBS was around 24 percentage points higher than that of manufacturing as a whole, and around 33 points higher than that of the total market economy. However, they did not include the main P-KIBS (law, accountancy) and C-KIBS (advertising), which were included by Hipp et al. (2015)[1]. These authors reported that KIBS and manufacturing overall presented similar rates of innovative firms – around 37 per cent of respondents within each claimed that their firm was innovation-active. This share varied widely across KIBS activities, with less than 20 per cent of consultancy and labour recruitment services portraying themselves as such. This contrasted with more than 50 per cent of various computer-related and over 77 per cent of R&D service firms.

While we have little data on, for example, the precise new technologies or marketing arrangements that are being introduced, the CIS does differentiate between four main types of innovation: product, process and marketing and organisational innovations. Do KIBS specialise on particular types of innovation as compared to other sectors? Howells et al. (2004) found over a third of service firms, as compared to fewer than 6 per cent of manufacturers, claimed their main innovations were organisational. However, their sample is small and biased towards larger firms. Using more representative CIS data, Schmidt and Rammer (2007, German data) and Miles (2008, UK data) found less striking trends, comparing shares of firms in different sectors that report introducing technological (product and process) and non-technological (organisational and marketing) innovations[2]. Both studies reported that sectors pursuing more of one type of innovation were also likely to pursue more of the other type. There was some tendency for services in general pursue organisational innovations more extensively than manufacturers; but this was far from suggesting an overwhelming, let alone exclusive, focus on such innovations.

We can examine these phenomena using more recent CIS data. Here, we contrast UK and German data for 2014. There is much similarity in terms of relative innovation propensities across industries, but some telling differences emerge. We illustrate these results in Figures 1 (a) and (b), and Figure 2(a)-(d), displaying data on innovation propensity for different sectors for Germany and the UK.

Starting with Figure 1, in Germany, while the correlations are not absolutely perfect, it is apparent that there is a tendency for sectors that do or do not engage strongly in one class of innovation, to behave in the same way in the other class of innovation, too. This pattern was also apparent in the earlier studies for Germany and the UK. But now it is much weaker in the UK, because levels of organisational innovation are much less dispersed across sectors: roughly half of UK firms in most sectors feature it, while the incidence across German sectors ranges from <1/3 to >4/5. This variation is hard to explain: possibly the UK’s high exposure to the global financial crisis has prompted a wave of organisational change in its industries; or possibly the data reflect the rise of the “gig economy”, with increases in the number of people working on zero hour contracts or as outsourced “freelance” workers[3].

The sectoral pattern of “technological” innovation is fairly similar across the two countries. Two T-KIBS – J62 (Computer programming, consultancy, etc.) and M72 (R&D) – are prominent in both. The P-KIBS M69 (Legal; Accounting), M70 (Head offices; management consultancy), M71 (Architecture, Engineering, Testing, etc.) and M73 (Advertising; market research) are relatively low in both. In relation to “nontechnological” innovation, M73 is among the top three in each country, and M69 is in the bottom two.

There is little sign here that the eight KIBS sectors place special emphasis on organisational innovations. In both countries the average for manufacturing (section C), is located fairly centrally within the swarm of KIBS. In the UK three T-KIBS - J62, J63 (Information services) and M72 - fall on the “technological” side of the diagonal, but five KIBS sectors did so in Germany, including again J62 and M72, but also M70, M71 and M74 (Other KIBS). Nontechnological innovation is important for many KIBS, but is not outstandingly predominant. There is thus some indication that technological innovation is more prevalent among T-KIBS, but hardly to an overwhelming extent. Nontechnological innovation seems to constitute a large share of innovative activity in P- and C-KIBS.

Looking in more detail at the four types of innovation distinguished in CIS, and plotting sectoral data from one country against the other [Figure 2 (a)-(d)], reveals some interesting details. Visual inspection of the scatter plots suggests that the innovation rates in the two countries follow roughly similar sectoral patterns[4]. Sectors with higher rates of one type of innovation in one country will tend also have higher rates in the other country. The major exception involves product innovation, where J63 emerges as an outlier, with a substantially higher innovation propensity in the UK than in Germany. J63 may also be an outlier in process innovation, again with a greater propensity reported in the UK. Less markedly, M74 shows more organisational innovation in the UK than would be expected from the general trend. Organisational innovations actually have a less pronounced trend than the other innovation types. This reflects the feature that other than M74, UK firms adopting these innovations range narrowly between 30 per cent and 50 per cent, while Germany’s range extends from 20 to 70 per cent. The sectoral propensities to undertake marketing innovations also display a notably smaller range in the UK than in Germany.

KIBS generally emerge as among the most innovative sectors for all types of innovation. Overall, the P-KIBS seem less liable to undertake most forms of innovation – not just technological ones – than are T-KIBS. (Occasionally one of other set of KIBS has a low innovation propensity, especially M69 for product and process innovations, and M70, M73 and M74 for process innovations.) Some T-KIBS are generally at the extreme of innovation-intensity. For product innovations, this applies to J63 only in the UK; J62 and M72 are high in both countries. J63 is a leader in process innovation in both countries, and M72 is in Germany. J62, J63 and M73 are forerunners in marketing innovations. M74 is outstanding in organisational innovations in the UK, with J63, M73 and M72 performing strongly in both countries (though there is little to differentiate among sectors in the UK).

In general, German firms are more prone to report most forms of innovation than UK firms (this is most debateable in the case of organisational innovations). It is not uncommon to find double the share of German firms in a sector reporting innovation than for UK firms. It is unclear how far this reflects actual differences in innovative behaviour (and if so, how far these are related to the specific conjuncture, as opposed to being deep-rooted differences)[5], cultural differences in self-reporting (British reserve?) or varying interpretations of “innovation” in English versus German. Cross-national comparisons using this sort of survey question should always be treated with caution.

CIS data can be used to address the issue of firms that only engage in one or other form of innovation (Figure 3). In no KIBS sectors did the shares reporting only organisational and marketing innovation, in either country, reach the levels (Howells et al., 2004) reported for services as a whole. Fewer than 25 per cent of firms do so, with the exceptions in Germany of M73, and M69 in the UK. In both countries, these two sectors – and section H, financial services, in the UK – had relatively high levels reporting only “nontechnological” innovation, with very low levels reporting only “technological” innovations. J62 and M72, and manufacturing industry, in contrast, were more prone, in both countries, to report the reverse pattern. More sectors feature in the technology-only category in Germany, more in the non-technology only category in the UK, again supporting the idea that the two economies differ substantially in innovation orientations. In addition, in both countries there is an evident tendency for those sectors that feature more firms specialising in one type of innovation only to feature fewer firms specialising in the other form of innovation. (It is in principle possible for both specialisms to be prominent if the sector had few firms that combined the two types of innovation – sectors in the upper right-hand areas of Figure 3 are heading in this direction, but there are always more firms undertaking combined innovation than pursuing the specialisms).

It is surprising that the KIBS literature, and indeed that on service innovation in general, has not paid more attention to the patterns displayed across types of innovation. There have been many studies examining such features of innovation in KIBS as Intellectual Property protection efforts, and particularly sources of innovative ideas and patterns of collaboration (Examples include Nählinder (2005), Mansury and Love (2008), Tödtling et al. (2009); Gotsch et al. (2011), Schricke et al. (2012), Doloreux and Shearmur (2013), Shearmur and Doloreux (2013), Bolisani et al. (2014), Asikainen (2015) – see also the extensive bibliographic studies cited in f2018)[6]. Probably the most extensive survey-based (CIS) study is that of Hipp et al. (2015), which covers almost 10,000 firms across 27 EU member states confirming high levels of innovation in KIBS, and also high levels of collaboration with others (mainly other domestic firms) – though the latter varies considerably by sector.

Sectoral generalisations can be potentially misleading, though they do provide useful insights into explanatory factors. The important point is that there are almost always variations within sectors, as illustrated by Hipp and Grupp (2005), using firm-level data. Surveying German service firms, these authors identified four innovation styles – knowledge-intensive, network-intensive, scale-intensive and external innovation-intensive. Though one or other style would usually predominate in any one sector, some firms would be using other styles. They found that the knowledge-intensive style was, as expected, very prominent among German T-KIBS – software, technical and R&D services – and in other business services and “other services” (There were also firms with knowledge-intensive characteristics in trade and transport, though rather few in finance). The network-based style was also important in these KIBS industries, along with the supplier-dominated style for technical and R&D services.

Hipp et al. (2015) also discuss patterns of innovation expenditure in KIBS, finding that KIBS do undertake R&D but they typically spend less on equipment purchases than manufacturers. We can update their results with recent CIS data for the UK and Germany (Figure 4)[7].

These data confirm that R&D constitutes a large share of innovation expenditure (and not only in these KIBS), though there are variations across countries. In the case of M72 (R&D services), R&D for one’s own innovation is a high share of innovation expenditure across all countries studied (this applies to expenditure on external R&D as well as in-house spend). “Other” expenditures – which could include training and market preparation – are important for M69 in Germany. Expenditure on equipment is rather important in some KIBS – especially in M69 (legal & accounting) in the UK; although Germany is far less active in such purchasing here in this sector, its investment shares in M70 (consultancy), M71 (architecture and engineering) and M74 (other KIBS) are quite substantial and exceed those of the UK. Rather surprisingly, this is less the case for the two IT-related KIBS. Again we find substantial variations across countries, reinforcing the warning in f2018 that data from specific countries cannot be assumed to be representative of KIBS worldwide (or even Europe-wide). Expenditures may be focussed on helping the firm, its employees, or its customers adapt to innovations or on developing the knowledge and facilities required to create and implement innovations. Much research has focussed on the latter issues.

What about the effects of innovation? Some studies consider the implications of innovation or lack of innovation for firm survival and growth (Baldwin, 1995; Cefis and Marsili, 2005); the general conclusion is that innovators may well be more successful than non-innovators, even if success is determined by a great many factors, some highly contingent ones. Innovation-related changes – or potential changes – in work and employment are currently the source of much debate, and we might expect KIBS – as vanguard innovators in many respects – to provide significant witness to these implications. However, the very features that make KIBS so distinctive – notably the high share of Professional workers – may also limit or shape innovation and/or its implications for working life. We shall thus first discuss the more general literature on current and future relations between innovation and working life, before turning to examination of some key features of KIBS occupations and employment. Finally, we will consider discussion of the future prospects of work and employment in KIBS.

## Innovation, employment and work

There has been an extensive literature on innovation (especially technological change) and employment, in which one theme is recurrent – fears of mass job loss resulting from automation and/or migration of whole industries or particular jobs to other parts of the world. In the 1960s, fears in Western countries that jobs would be lost in manufacturing were countered by evidence of job growth in service industries, but more recently commentators have been making the case that not only routine services, but also many KIBS, might face similar threats through new generations of automation and offshoring. A second line of analysis has concerned deskilling or the degradation of work: the argument is that technological and organisational change are being shaped so as to reduce the scope for discretion and creativity at work, even in professional jobs. A more recent body of arguments asserts that current technological changes are fostering a polarisation of the workforce and thus social and economic inequalities within societies[8].

We will focus on empirical research, as it concerns KIBS in the following sections, though to do so we will also need to examine more general trends and discuss some of the most prominent forecasts articulated around these themes. In some cases, the literature is more concerned with occupations than with industries, and we can use data more relevant to KISA (knowledge-intensive service activities, such as those in professional occupations). As we shall see, professional occupations and KISA work predominate in KIBS.

### Employment in knowledge-intensive business services: key issues

When it comes to overall employment levels, the research reviewed in f2005 and f2018 makes it clear that in most countries KIBS industries have been expanding more rapidly, in terms of employment as well as of output, than has the economy as a whole (or than most other service sectors). Thus, even though KIBS sectors report higher rates of innovation than most others, innovation does not seem to have presented a great challenge to KIBS’ employment. It might be argued that the amounts of innovation undertaken in KIBS are less dramatic than those in some other sectors, and thus have less evident employment impact. Comparing information on expenditure levels per employee on innovation might be one fruitful way of exploring this in more depth, and as noted earlier[9], such ratios are considerably lower for P-KIBS than for manufacturing and T-KIBS (in German data, at least).

Most forecasts from national and international employment, skills and training agencies apply econometric modelling based on these trends, and thus anticipate continued growth for KIBS. For example, Cedefop (2016) presents authoritative forecasts for the EU28[10]. In its baseline scenario for the decade to 2025, jobs are forecast to be lost in the primary sector, utilities and manufacturing. Services – especially “ business and other services” are expected to show most employment growth the (1.1 per cent per annum, compared to 0.4 per cent for transport and distribution, and 0.3 per cent for nonmarketed services; construction is at 0.1 per cent, while other sectors are negative)[11]. Cedefop expects new technologies like automation to replace much routine and data processing work, both blue- and white-collar, but to create demand for specialised skills and high qualifications.

This reassuring picture is challenged by a number of recent writers, notably the substantial book by Susskind and Susskind (2015). They examine eight professions, and conclude that, despite many variations in detail, all are liable to be transformed in coming years, and a number of common trends that are already apparent across professions. These are:

• shifts from bespoke to more standardised services;

• bypassing of traditional gatekeepers (e.g. the rise of new online services, including new brokers and intermediaries like comparison and advice sites, “communities of experience” where users can exchange knowledge, etc.);

• movement from reactive to proactive services (anticipating client needs, offering preventative as well as problem-solving inputs); and

• efforts to provide “more-for-less” as clients face cost pressures.

Cost pressures are indeed identified as major drivers for these developments, along with technological advance (artificial intelligence, big data analytics, pervasive computing, etc.) These pressures vary across tasks (components of the professionals’ responsibilities) and across professions, so jobs are more likely to be transformed than simply abolished, though the tenor of the analysis suggests considerable job loss. The Susskinds pay little attention to possible social drivers – for example, levels of trust in expertise, desire for autonomy and rebellion against perceived powerlessness, concern about the security and use of personal and confidential data and debates about knowledge as a public good (this idea could have wide resonance, though is most prominent at present for education, health and perhaps legal professions).Whether social drivers such as these are more liable to slow or speed the transformation of professions’ roles (and thus KIBS) is unclear.

Many studies have attempted to identify how far different professions might be at risk from the new technologies. For instance, Frey and Osborne (2013) estimated that about 47 per cent of American workers had a job with a risk of automation: some jobs in KIBS (e.g. legal secretaries; bookkeeping, accounting and auditing clerks) faced almost 100 per cent probability of computerisation (Berger and Frey, 2016; Frey and Osborne, 2017) Unlike some more alarmist accounts, these authors recognise that just because a job can in theory be automated, it will not necessarily be. The Susskinds occasionally cite statistics, but do not spell out employment/job loss forecasts, with much of their argument based on amassing numerous specific exemplars and instances of trends. So far, we see that high levels of innovation in KIBS coexist with growing employment in these industries - their implicit argument is that revolutionary new technologies (and the social circumstances that are conducive to their use) imply change of a qualitatively distinct order from longer-term experience. Hopefully, their provocative study will stimulate substantial research examining the trends and drivers that they identify. The sorts of data analytics discussed in the book could actually be used to explore issues like the growth of new online intermediaries and self-help facilities (indeed, it would be surprising if they were not studied by those concerned with marketing of services).

### Outsourcing and offshoring

Several employment-related topics can be examined with readily available statistics, First let us consider outsourcing (how far service activities are contracted to KIBS firms, as opposed to being undertaken by user organisations themselves) and offshoring (how far activities are moved to other countries). The users of KIBS may outsource longstanding knowledge requirements or commission inputs to confront novel requirements – alternatively, they may seek to develop internal knowledge capacities. Actually, as knowledge is required to effectively source and apply external knowledge inputs (Sundbo and Gallouj, 2000), KIBS users are liable to face limits to how far they can completely contract out knowledge requirements to external suppliers. The business case for insourcing or outsourcing knowledge resources will vary across firms and situations; as we see below, there are big differences across different areas of knowledge.

It is possible to examine the location of at least some KISA by examining detailed occupation-by-industry data. Here again US statistics provide scope to examine trends (Table I). There are striking variations in the extent to which the KIBS sectors are the major employers of the relevant professions, from around 10 per cent in the case of design to around 80 per cent in that of architecture. With the exception of lawyers, the tendency has been for a growing shares of the profession to be based in the specialised KIBS sector. The time-span considered is brief, and more data points are required to establish a trend. However, it does not look as if KIBS generally are seeing their key roles being taken over by user industries, in the USA at least. While other data are required to address the Susskinds’ points about new online sources of support, new sorts of brokerage and the like, there does not seem here to be a trend toward business user insourcing of KIBS-type inputs.

The growth in KIBS jobs experienced in Western countries could be limited by the offshoring of much professional work. This issue has been frequently addressed, with several Bureau of Labor Statistics (BLS) and OECD studies providing quantified. BLS paved the way in developing methods for estimating how far different jobs might be amenable to offshoring. Moncarz et al. (2008), who also review earlier studies, elaborate the BLS approach. After filtering out the many service jobs that rely on face-to-face interaction or the service worker’s physical presence, 160 jobs remained. Experts rated these in terms of four supposed determinants of susceptibility to offshoring: requirements for

• inputs and outputs that can/cannot travel easily across long distances;

• interaction with other types of workers;

• knowledge of the social or cultural idiosyncrasies of the target market; and

• routine/non-routine work.

Overall, the susceptible jobs (and often those with medium levels of susceptibility) were occupations that had been increasing more rapidly in terms of numbers and wages than jobs as a whole. Some of the main occupations associated with KIBS work were seen to be relatively open to offshoring – many occupations within accountancy, architecture, computer work; others were less so – for example, legal occupations. Most of the affected occupations were ranked in the middle level of susceptibility.

Other studies have tried to put numbers onto the shares of employees that might face threats from offshoring of their jobs. Sophisticated analysis the offshoring potential of various types of jobs has recently been undertaken using data on a wide range of characteristics of occupations in Germany, based not on expert judgements but on survey data. Brändle and Koch (2014), applying principal components analysis to these data, develop indicators differentiating between outsourcing and offshoring potentials[12]. They examine results for the two indicators, aggregated in terms of occupations, sets of tasks and industrial sectors. In terms of occupations, highest offshorability is mostly associated with mid‐skill routine jobs – in line with accounts of job polarisation considered below. Some KISA professions (e.g. chemical engineers, and architects and civil engineers) are seen as having low outsourcing potential[13]. Considering tasks, least offshoring potential was displayed by such KISA as research, data processing and organising tasks. At the sectoral level, business services (they do not distinguish KIBS) were relatively low in terms of offshoring potential. This sounds like good news for KIBS, though note that Brändle (2015) examined the link between the offshorability of businesses’ workforce (measured by the indicators of potential for offshoring work discussed above) and the firms’ actual offshoring behaviour, he found a negative link – German firms with more offshoring potential were less actually likely to offshore parts of their production processes. At the very least, such a surprising result makes us chary of reading too much into forecasts based on assessments of jobs’ vulnerability[14].

Sectoral levels of offshoring have been addressed in studies of trade. f2018 (cf. its Figure 12) reported that over the period 2001-2015, for the EU, Japan, the USA and also for the BRICS countries taken as a group, the share of KIBS in their exports was growing[15]. While the growth in BRICS was relatively fast, the share of trade in these economies remained below the level for the other regions (with the exception of Japan). Sako (2014) mainly focuses on motivations of users to offshore professional services, and on the problems of conducting professional services remotely – especially those requiring face-to-face contact. She reports data on both exports and imports of business, professional, and technical services for the EU and USA, for 2010. For the USA, exports are higher than imports for all sectors considered. For the EU27, this is the case for Legal, accounting, management and public relations; architectural, engineering and other technical services (the largest group of sectors by value); and also agricultural, mining and other on-site processing services (the smallest group). EU27 imports were greater than exports for advertising, market research and public opinion polling and for R&D services (respectively, the second and third largest groups by value). Some of the EU trade deficit is probably accounted for by imports from the US, but other countries are probably also involved.

Liu and Trefler (2008) focussed on outsourcing (imports) and insourcing (exports) of services from the US to China and India, over the period 1996-2006. The largest traded category was Business, professional, and technical services (BPT - $33,637m of exports in 2005,$13,490m imports); followed by financial services, insurance (the only large category to show more imports than exports) and telecommunications. Within BPT, R&D services and advertising were the only subgroups with more imports than exports; and larger subgroups like IT, legal and architecture and engineering services show substantially more exports than imports. Overall, though, BPT imports from China and India grew faster than did exports to these countries. Liu and Trefler conclude that these trends on labour markets have actually had small and mainly positive impacts in terms of US professional jobs; only low-skill white-collar jobs suffer negative impacts. The prospects of more professional work being offshored from the richer countries (where KIBS are most prominent) to newly industrialising countries is the subject of continued speculation. Concern may be fuelled by experience in the West will be unfamiliar with telephone help lines and telesales calls originating from the Indian subcontinent, and of Chinese-originated high-tech manufactured goods. But much current offshoring is between industrialised countries, and from these to other regions; and indigenous KIBS should contribute the development of emerging economies, and potentially add diversity to at least some KIBS offerings. While there can be impacts on some professional occupations in the West, the threat seems to be overstated, especially when compared to the challenges of new technologies. These challenges may be more important, furthermore, in terms of job content than of job numbers.

### Occupations and occupational structures in knowledge-intensive business services

The ideas of deskilling and job degradation involve the idea that higher-skill jobs are being replaced by lower-skill ones, and/or that the skills required of particular occupations are becoming attenuated. Braverman (1998) argued that this derived from managerial interest in having a higher level of control over the workforce (together with a lower wage bill!) Data for the economy at large suggest that while some skilled work is replaced by less skilled work, there has been a tendency for Western economies to generate more skilled jobs over the long term, however. (There is said to be a “skills bias” in technological change). However, many such jobs are in KIBS, and few empirical studies or forecasts have addressed deskilling within KIBS work and KIBS-type occupations (KISA). We can examine several phenomena that are related to this topic, illuminating features of KIBS evolution beyond simple counts of those employed.

Forecasts of employment levels often distinguish between different jobs, and some studies go on to discuss changes in the skills required of jobs. Cedefop (2016) anticipates for continuing displacement of routine and data processing tasks and growing demand for specialised skills and high qualifications. Cedefop uses the ISCO (International Standard Classification of Occupations) framework, which features ten main categories of job. Three of these mainly involve KISA, and are prominent in KIBS (Martinez-Fernandez et al., 2011): ISCO categories 1 to 3, respectively, Managers, Professionals and Associate Professionals/Technicians. (ISCO1 is somewhat problematic, roughly a quarter of those in this category in the UK in 2017 are self-employed in all kinds of trade, while ISCO1 also covers a wide range of roles, extending to senior judges and politicians and executives of large corporations.) ISCO4 covers information-intensive but mainly more routine Administrative, Secretarial and Clerical work.

Cedefop (2016) also forecasts that occupational groups encompassing ISCO 1, 2 and 3 will grow over the decade to 2025. The job opportunities will be around 16m for ISCO1 positions (employment growth of 0.9 per cent p.a.), over 43m for ISCO2 (0.8 per cent), and over 41m for ISCO3 (1.0 per cent). The great majority of the job openings involve the replacement of workers who are retiring or otherwise leaving the occupation, rather than involving the creation of new positions – fewer than 10m of the positions are new, rather than replacement, jobs. ISCO2 (professional) jobs are particularly demanding of graduate-level education, with about 80 per cent of occupants in 2025 forecast to have attained this, as opposed to less than 40 per cent of the workforce as a whole (over 60 per cent for sISCO1, over 0 per cent for ISCO2). Cedefop expects qualifications to become more necessary for these positions in the future, this runs counter to the deskilling thesis; though it could be argued that higher levels of University education in the society at large are simply being reflected in the employment projections[16]. Over the whole economy, some polarisation is projected: ISCO4 jobs (clerks) are expected to decline (by 0.3 per cent p.a.), as most other categories are expected to grow. (The exceptions are mainly medium-skill physical jobs: ISCO6, skilled agricultural and fishery workers, – 1.3 per cent; ISCO7, craft and related trades workers, −0.6 per cent; and ISCO7, plant and machine operators and assemblers, −0.2 per cent).

Such forecasts are produced by combining basic econometric models with extrapolations concerning the occupational structure of industries; in general, such studies yield bright prospects for KIBS and for professional work. The same is true of several consultancy studies examining KIBS. Thus, EMCC (2006a) forecast that, in general, KIBS will require increasing numbers of highly educated, reliable and innovation-focus employees, and that educational institutions and training programmes be providing relevant technical and scientific capabilities, networking and social skills, as well as targeted business know-how. The initiative for such development may well stem from large companies and/or industry associations. (It is noted that KIBS are liable to cluster in particular locations, with implications for Universities and other bodies in these sites.) The mobility of KIBS Professionals within the EU would be assisted by the development of more focussed accreditation, qualification and training schemes.

Some studies of the impact on professional occupations of offshoring, organisational development and new technologies will be complex. For example, Winterton et al. (2000) reviewed future skill needs of managers in the light of changing socio-economic, political and technological trends in the UK, France and the USA. The study asserted that traditional management structures are being displaced by more horizontal structures involving cross-functional core process teams featuring initiatives to build involvement, participation, team working and autonomy. Future skill needs of managers would include knowledge-based technical specialisms in addition to generic management competences and more group-orientated leadership styles based on trust and collaborative employment relationships.

A rare study of occupational trends in KIBS is a “mapping” report (EMCC, 2006b). Both pieces of work mainly focus on consultancy-type KIBS, but this mapping study makes several points of more general relevance. Among the trends in skill requirements and work organisation that are discussed, the study indicates that many KIBS companies use more flexible staffing models that culminate in a leaner workforce. Many KIBS are expanding their service offerings to provide a fuller range of services, which means recruitment of staff from a wider variety of backgrounds, and the development of more vocational skills (to develop practical solutions) and social and network-based skills (for collaborative working arrangements with clients). Different KIBS are liable to follow different development paths; the most likely strategy for the consultancy industry is close collaboration between the KIBS organisation and the client firm, as KIBS establish high levels of client-firm specific knowledge and “deliver whole packages of improvement”. KIBS work is therefore increasingly consultative, with the company working in partnership with clients to solve their problems, requiring high levels of competence and trust. Purely academic skills and credentials are less important than a combination of communication skills and experience in solving business problems (required to efficiently produce positive client outcomes). However, within many KIBS subsectors there are pressures to standardise services to benefit from economies of scale (e.g. using modular solutions that can be customised to client needs), which may lead to more distant relationships with client companies.

We noted earlier that Susskind and Susskind (2015) foresee more dramatic changes in professional work and skill requirements, reflecting changing demand and new technological capabilities. Many tasks may be automated (or at least supplemented by “decision support” systems). But new skills and tasks will arise, associated, for example, with using advanced IT systems and big data, communicating via online systems – including social media – and the like. The successful firms and individuals are liable to be those that recognise that some of their traditional tasks may be better performed by machines or online communities, and that they accordingly can carve out new roles for themselves. The Susskinds refrain from quantitative forecasts, but present several distinct scenarios – which could well coexist in various ways, and emerge at different rates – as professionals and professional organisations adapt to the challenges and seize the opportunities that are encountered.

Again, some notes of warning are sounded in recent research. Henseke et al. (2018) suggest that new generations of IT are less skill-demanding than was earlier the case (which corresponds with expectations that we would draw from product cycle theories as well as everyday experience). Henseke et al. (2018), present data from UK Skills and Employment Surveys, conducted over the present century, indicating that in 2001 around 25 per cent of survey respondents stated that additional computing skills would enable them to do their job much better, but only 12 per cent did so in 2017. It may, then, be unsafe to extrapolate from historic technological change to future prospects; the “skill bias” of technological change may be overstated.

Analyses of earlier rounds of the UK Skills Survey have indicated another relevant trend, which could be subject to more exploration. One implication of this may be that “decision support” is frequently a matter of “decision removal” from at least some KISA workers – though we do not know how far the UK experience is paralleled in other Western economies. The Surveys examined the extent to which workers are able to exercise skills of discretion and decision making in their jobs. Felstead et al. (2013) examine employee views from 1992 to 2012 about the amount of discretion they have over daily work activities. Discretion was seen to decline for all occupation groups over this period, especially in the 1990s (trends were much flatter in the present century). Managers retained higher scores than Professionals and Associated Professionals, who began with relatively high scores but then declined to a level near the average; Administrative and Secretarial workers moved from a near-average to below-average level. Inanc et al. (2013) use these Surveys to examine another aspect of working life: they found a trend of work intensification over this period, associated with technological change, which the authors thus dub “effort-biased”. More qualified employees were more likely to report having to work very hard. These studies also provide some evidence as to the actual skills employed at work. Briefly, Felstead et al. (2007) reported that overall occupations in the late 1990s and early 2000s experienced greater requirements for most generic skills (planning skills, for example) – contrary to the deskilling account; this was mainly involved lower status occupations; Professionals mainly reported stable usage of skills, with some declining requirements (e.g. for number skills), though still reported relatively higher skill usage[17]. Another source of evidence indicating change in the task content of jobs comes from European-wide survey work reported in Eurofound (2016).

Between 2000 and 2010, the share of the workforce in jobs that feature high levels of routine work decreased; but over the same period, many jobs acquired more routine elements. Increases in routinisation related to repetitive work were not found in our KISA occupations (though they were present for clerical workers); increases in routinisation related to standardisation were also found for managers and professionals; and autonomy was reported to decrease for some managers, professionals and associate professionals (Eurofound, 2016, p. 68)[18].

### Occupational structures: professional, associate and clerical workers

The Eurofound analysis requires examining both change in the task content of jobs and change in the occupational structure of employment. One approach to studying change in KIBS work involves examining trends in occupational structures – future work should examine also the task content of occupations. Relatively few studies have examined the composition of employment in KIBS. (Consoli and Elche, 2013, are a major exception.) f2018 presented data on the gender distribution and educational qualifications of the workforce, the size distribution of firms, in various KIBS sectors. However, it did not examine the structure of employment in occupational terms, which is highly relevant for considering the future of employment in KIBS.

Let us begin with British census data for 2017 that allows us to examine the occupational structure of different industries[19]. Figure 5 displays these data for KIBS as a whole and for the three main classes of KIBS discussed in f2018. These are KIBS based on traditional professions (P-KIBS), those related to science, engineering and new technologies (T-KIBS) and those focussed on cultural and creative activities (C-KIBS)[20]. The data confirm the anticipated distinctiveness of KIBS compared to the whole economy. KIBS are not only particularly graduate-intensive and innovative; they are KISA-intensive. We now note that ISCO2, Professionals, constitute the largest occupational category in KIBS in general, with this share being more than double that of the total economy (England and Wales). ISCO3 (Associate Professionals and Technicians) and ISCO4 (Clerical Support) are the second and third most prevalent jobs in KIBS; ISCO2-4 together account for roughly 80 per cent of KIBS employment, around twice the share featured in the economy in general[21].

The data show, however, some differences between the major groups of KIBS. C-KIBS feature more sales workers; T-KIBS display the highest share of Professionals, and low levels of administration/secretarial occupations, while P-KIBS display a very large share of the latter.

It is possible to delve deeper into the occupational classifications from the published data; but this reveals mostly rather predictable tendencies – S&T Professionals tend to feature in T-KIBS, Business, Media and Public Service Professionals in P-KIBS, and the like. It is noteworthy that slightly more employees in ISCO4 are in administrative roles, rather than secretarial and related work. Examining individual KIBS industries in more detail again tends to yield predictable results, though few intriguing features emerge, for instance, a relatively high proportion of ISCO1 in management consultancy (reflective of small businesses?) but also in the advertising sector; relatively many sales workers in testing services.

Analysis of employment prospects for KIBS will need to be attentive not only to variations among KIBS, and but also variations across the KIBS workforce, in terms of types of work undertaken. Professionals may be expected to undertake more tasks traditionally performed by support workers – using keyboards is an obvious example. However, also, more of the work traditionally associated with professionals may be delegated to support workers: “paraprofessionals” have been enlisted to perform many more routine tasks in some knowledge-intensive work. We can examine the ratios of particular occupations in KIBS sectors to each other: for example, how far are professional occupations in KIBS receiving support from ISCO3 and 4 staff, and how does this change over time? It may be that support workers are becoming more efficient, and each worker can provide more support by the application of new technologies and other changes. Thus, the ratios provide only a first glimpse into the interplay of the underlying processes described above.

Trends in the occupational structure of KIBS can be obscured by changes in both industrial and occupational classifications, which can make it laborious to construct time-series data extending over periods of decades. However, in many countries, alongside the increasing share of professional workers in the labour force is a declining share of secretarial and clerical workers. This applies to KIBS and the wider economy. Thus, the Australian Bureau of Statistics publishes data spanning several decades on the occupational structure of Professional, Scientific and Technical Services (this includes computer services). As Figures 6 and 7 show, Clerical and Administrative Workers decrease their share, Professionals increase theirs. Indeed, the number of clerical personnel providing support for each professional is more than halved between the 1980s and recent years (increased productivity means that this may not be reflected in terms of level or quality of service). From a situation of almost one Clerical worker per Professional in the 1980s, these Australian KIBS now feature around one for every three. A similar, if less marked, trend is apparent in terms of Technician/Associate Professional support (Figures 6 and 7).

Shorter time-series for other countries is available. Figure 8 presents data for the EU15, and France, Germany and the UK, over about a decade. The trends are similar, though the ratios vary considerably across countries. In particular, there are considerably more associate workers for each German professional worker, across the economy and also within the KIBS sectors, than in the UK. The overall EU15 ratio is more or less between these two. There is a sharp drop in the ratio in the wake of the 2008 crisis; it recovers to some extent for the whole economy in Germany, and for KIBS at the EU15 level, but otherwise seems to return to a slow but steady decline.

The ratio of Clerical workers to Professionals is much lower in the total economy and for KIBS than that for Associate Professional/Technical workers. There is again a sudden drop post-2008, especially for the UK (which had begun the time-series with the highest ratio of Clericals to Professionals, but which plunged to the bottom). One interpretation is that there was much less of a negative impact of the crisis on professional jobs (which at EU15 level, at least, continued to grow in number through the period), while the other occupations suffered some years of reductions in job numbers and/or growth rates[22]. The clear trend, even over this fairly short period, resembles that of the Australian data: professionals’ share of the workforce grows compared to that of the workforce supporting their activities.

The decline in clerical employment is likely to reflect the near-universal adoption of word processing and related office equipment. Managers, professionals and, indeed, many other workers regularly enter data themselves on personal (and increasingly, mobile) devices, fulfilling some of the fears about the future of secretarial work expressed in the 1980s, during the onset of the “microelectronics revolution”.

A similar trend appears to apply to the relationship between Professionals (ISCO2) and Associate Professionals (ISCO3). Though there may be a tendency to create more work for ISCO3 by shifting more routine elements of the Professional job onto Associates, these tasks may be more subject to automation or rationalisation of one sort or another. Fewer Associates may be required to perform these processes – or professionals might use technology to accomplish tasks previously allocated to Associates. This looks likely to underpin the decreasing ratio of Associates to Professionals. However, how does this ratio look across different KIBS sectors?

More detailed occupational information is available on subcategories of Associate, and Figure 9 demonstrates that for England and Wales, at least, in 2017, all KIBS sectors have some employees belonging to each of the subcategories of Associates (and Professionals) distinguished in the statistics. Not surprisingly, Associate Professionals tend to be concentrated in those sectors related to their types of knowledge. Thus, S&T associates are more prevalent in T-KIBS, while Culture/Media/Sports Associates form a marked share of C-KIBS but are relatively uncommon elsewhere. Business-related Associates are substantial in both C- and P-KIBS, and are actually more numerous in T-KIBS than are S&T Associates (7.8 per cent versus 7.2 per cent shares). Requirements for technical and business skills are spread widely across KIBS; “cultural” skills are much more concentrated in specialist services. In terms of future trends, the impacts of automation are liable to vary across the different tasks associated with these roles, including the demands for creative initiative and interpersonal contact that they involve.

It is not clear whether the declining ratio of Associates to Professionals reflects Professionals taking on themselves some of the work accomplished by Associates, or of Associates becoming more productive (possibly even taking on more of the work of Professionals). Answering this requires examination of the work conducted by Associate Professionals and Technicians. Rodgers and Waters (2001) present one of the few studies to address this topic, mainly focussing on P-KIBS. They examined five types of Business and Public Service Associate Professionals: insurance underwriters, legal Associate Professionals (legal executives and barristers’ clerks), personnel officers (including recruitment consultants), market researchers and estate agents. Despite different skill needs, entry routes and skill development strategies, the study was able to identify crosscutting features and trends. The study provides much detail on specific skill requirements, but it identified three common skill mixes. Associate Professional occupations were mostly:

• “traditional”– requiring a high level of technical skills with above-average generic skills and well-developed personal attributes;

• “generic” – requiring high-level generic skills and personal attributes but relatively low levels of technical skills (typically, the skills are largely transferable, the jobs have lower entry requirements, and employee turnover is higher); or

• “transitional” – requiring an average level of technical skills, but high-level generic skills and well-developed personal attributes.

Jobs in this third group may be undergoing restructuring, incorporating additional job tasks. A broadening of the roles was evident across all three role types, especially in the required level of technical knowledge and the need for interpersonal and “customer handling” skills (as users demand higher levels of “service”). If these results can be generalised, it appears that the work of Associates is indeed encompassing more tasks previously undertaken by Professionals, and thus is not subject to deskilling in any simplistic sense[23].

This study reminds us that the tasks and skills associated with jobs may be shaped by divisions not only of labour and available technologies within organisations but also the division of labour between organisations (including between KIBS firms and their customers), and also by changes in demand for service quality – and the emergence of new services. Process and organisational innovation are not the only factors shaping work content and employment structures. Service product innovation, and also service delivery innovation, may require new activities on the parts of KIBS staff. An obvious corollary is that new occupations may emerge, as well as the content of established occupations changing.

### Knowledge-intensive business services and polarisation trends

The deskilling debate of the 1980s has been overshadowed more recently by the emergence of a body of work on polarisation. This argues that, in the whole economy, and within firms and sectors, we may have neither an overall upgrading, nor an overall deskilling. Instead, recent years are characterised by what might be described as a disappearing “middle” as the share of jobs featuring mid-level skills (and wages) declines relative to those featuring higher or lower skill levels. One influential analysis is Autor’s “task-based” approach (Autor, 2013). Occupations are constituted from bundles of tasks. Some tasks (especially repetitive and routine physical and cognitive tasks) are more vulnerable to automation than others. Work that mainly features such tasks can be displaced by new IT systems. Two other classes of job are less easily automated:

1. jobs requiring problem-solving and interaction skills; and

2. those “low-skill” jobs that involve physical dexterity skills and adaptability beyond the capabilities of most current IT systems.

The recent rise of interest in polarisation has not so far been reflected in many forecasting studies, though we noted that Cedefop (2016) is projecting the trend. It is rarely addressed in studies of KIBS, but more general studies have proliferated (e.g. EU analyses from Eurofound, 2013a, 2015b). Some studies focus on a polarisation in wage levels, while others are less concerned with actual wages, instead contrasting high-, medium- and low-wage occupations, or with high-, medium- and low-skill occupations.

Eurofound (2015b) is able to present data spanning more than four decades, on occupations grouped by wage level, for six EU countries. Over this long span, the general picture was “consistent expansion of employment in high-paid jobs … contrasting with a very significant decrease in mid-paid and low-paid occupations across countries and periods” (p. 2). All countries lost mid-paid jobs as a result of deindustrialisation. The “more dynamic service sector” shaped overall structural change to a greater extent, in different ways across countries. Thus, growing public sector employment “was generally linked to the growth of relatively high-paid occupations, but in the UK, it also expanded the bottom quintiles after 1990, contributing to polarisation” (p. 2). Some countries display polarisation trends in more recent years: Germany and the UK from the early 1980s on, in Spain during recessions. In Ireland and Switzerland there was polarisation in the 1980s, though the general trend since the 1970s in these countries, and also in Spain and Sweden, has been one of upgrading. Data for 2011-2014 were available for a larger set of EU countries. In this period, across the EU, greatest growth was in well-paid jobs, with modest growth in low-paid jobs – and declining employment in middle-wage jobs. The authors termed this “asymmetrical polarisation”. Most countries displayed either upgrading or polarising trends; only a minority exhibited downgrading (growth restricted to lower-paid employment).

These patterns suggest that polarisation involves different combinations of automation and up- and de-skilling trends; variations in the mix at country level may reflect differences in industrial and organisational structures[24] (including the role of the public sector), and in industrial relations and policy climates. Since KIBS are among the most important areas of service sector expansion, trends in these industries are liable to be significant at the whole economy level. Furthermore, as highly innovative sectors, KIBS may be expected to be particularly susceptible to the impacts of innovation on skills, occupations and work organisation. How do these various claims about developments in working life look in the context of empirical evidence on KIBS?

We focus below on data for EU countries; very similar patterns of development are apparent in the US. These data concern mostly the past decade, which is most relevant to considering current technological and organisational changes. However, we should caution that the economic crisis that became manifest in 2008 promoted a period of low growth and job loss; more recent years show some recovery to different extents in different countries (the UK was particularly badly hit by both problems in the financial sector and in the recent Brexit concerns).

Much of the discussion of polarisation has concerned growing extremes in wage levels, but for the present discussion, analyses that examine trends as described by skill types are most relevant. Figure 10 draws on a classification of jobs into three categories, roughly following the ISCO sequence – thus the “High Skill” (HS) category consists mainly of ISCO categories 1 to 3. We can compare data for the economy as a whole with data for those KIBS in NACE section M, Professional and Scientific Services (unfortunately, this neglects the IT-related KIBS in section J). The EU15 economy as a whole shows a decline of around 5 percentage points in the share of “Medium Skill” (MS) jobs, with increases between 2 and 3 per cent for HS and Low Skill (LS) positions. Within Section M, HS positions predominate, and grow slightly; there are few LS positions, and their share increases a little; the share of MS jobs again decreases. KIBS’ polarisation pattern, apparent at the EU level, applies to France; in Germany KIBS’ LS share declines, in the UK KIBS’ LS share increases, but it declines across the whole economy. The MS share declines in all countries.

Because sectors grow at different rates, and KIBS have expanded relative to the economy as a whole in these countries, this polarisation pattern means different things when it comes to actual numbers (rather than shares) of jobs. The EU15, and the three countries, all display-marked polarisation in terms of jobs created across the whole economy. There is growth in HS and even more so in LS, with declining numbers of jobs in MS (Figure 11). Because of KIBS’ growth, numbers of MS jobs actually increase in section M in Germany and the UK, though not in France. In the UK numbers grew in all KIBS job categories, especially LS jobs; in France LS again dominated; but in Germany, numerical growth was relatively low and LS numbers declined[25].

Analyses of the US data (2003-2017) suggest that the pattern there is rather similar to the UK model, though the share of LS jobs in American Professional and Business Services category is much larger than in UK KIBS (section M). HS and LS shares increase both in the US economy at large and within these services. MS shares decrease in the US – again, exemplifying the polarisation pattern. Polarisation is more extreme in terms of the numbers of jobs involved. There are higher rates of increase of HS and LS jobs, and higher rates of decrease in MS jobs, than for the economy at large[26]. The national differences are liable to reflect a host of factors – the size of KIBS industries relative to that of the national economy, the share of different KIBS in the sector, the impact of economic crisis/recovery on KIBS’ users and varying cultural features of business organisation.

We can examine different types of KIBS in more detail using UK data, which can include the IT services not captured in section M. (As before, cautions about generalising from the situation of KIBS in one country to others apply.) For all three classes of KIBS in the UK, the pattern of polarisation was more pronounced than for the economy as a whole, over the period 2008-2017 – see Figure 12. The economy as a whole experienced substantial increase in the HS share of employment, decrease in the MS share, and a small decrease in the LS share. In contrast, KIBS display considerable growth in LS jobs – but from a very low base (only a few per cent of the KIBS workforce are LS). The LS share for KIBS as a whole more than doubled, except within T-KIBS. C-KIBS feature slight decreases in the shares of HS, as well as MS.

If we consider the actual numbers of jobs in each category, the overall employment expansion in KIBS means that across all KIBS categories, there is growth in HS numbers, lesser growth in MS numbers, and greatest growth in LS numbers, with a massive fivefold increase in C-KIBS (Figure 13). The rise of LS positions in C-KIBS is a phenomenon which will need further examination; but it is clear that the general picture for KIBS is a polarisation pattern, with (absolute or relative) shrinkage of mid-level jobs.

Polarisation is important for societies, in that it contributes to dualistic economies in which the life chances and interests of rich and poor threaten to diverge. Issues of social justice and political stability obviously arise: indeed, some commentators attribute the rise of populism and distrust of expert analysis to growing social inequality (as well as to pressures associated with globalisation). However, it also impedes social mobility, not least by reducing the scope for individuals to work their way up, through mid-level to high-level jobs. If this applies to KIBS, it will render their occupations less inclusive and diverse, restricting their social and cultural knowledge resources – and further fuelling suspicion of their contributions.

### New jobs?

In earlier decades, the expansion of service industries was seen as a source of new jobs to replace those that are lost by automation and offshoring of manufacturing. If these jobs are now threatened, where might new activities emerge? Many of the optimistic commentators will simply point to activities that are so far resistant to automation – especially creative activities – and say very little about their specificities. Even Susskind and Susskind (2015) speculate about new and emerging professional roles, including paraprofessionals (we did not see much evidence for this in statistics concerning Associate Professionals), new IT-related professions in data analysis and the like, and specialists in various new aspects of marketing and human relations. Other studies seek to determine new occupations where growth may be expected, and even to forecast the emergence of categories of work that have yet to materialise. The US O*NET team, for example, has developed systems for identifying, evaluating, and incorporating new and emerging occupations into its classification system, especially in high growth industries (O*NET, 2006) – though KIBS are not included among these, other than IT activities. Many of the 83 jobs considered in their work (Rounds et al., 2013) most are nonetheless Professional and Associate Professional roles (Bioinformatics Scientists, and Technicians; and Geospatial Information Systems Scientists and Technologists, etc.) that may be located in several KIBS industries. Ahlqvist (2003) used Delphi surveys to solicit expert opinion on new occupations related to key developments in IT, Biotechnology, and Nanotechnology/Materials technology (such as targeted medicines, sensors, intelligent materials), and several professions were deemed plausible in the coming decade, including Bioinformationist, Geoinformationist, Nanotechnology consultant and Visualisation specialist. KIBS were not a particular focus of the study, but some hints as to new KISA are provided. The tasks involved may exist, but Ahlqvist’s professional categories have not, in general, stabilised and reached wide acceptance.

These studies represent very different ways of exploring future professions, one very much grounded in empirical analysis of current trends, one based more on expert visualisation of future possibilities. Each suggests that there is plenty of opportunity for new jobs to arise: but neither gives ground for optimism that a huge number of such jobs will become rapidly available.

Yet the more apocalyptic projections of the impacts of the application of new technologies suggest that these impacts may be experienced in quite a short-time scale. While our own data analyses provide little support for arguments that extremely rapid change is already underway. But, given the efforts being put into applying technologies such as artificial intelligence and big data analysis to many types of professional work, it is likely that KIBS will undergo substantial change in the coming decades. In the past decades of the twentieth century, we saw mobile phones and a range of computer devices and systems helping to reshape many professional tasks. It is quite plausible that technologies now emerging will have a more profound significance, just as early generations of new IT have done for clerical and secretarial work (and probably also for that of associate professionals). The rapidity of these changes is the critical issue for thinking about job loss, and how far this is a threat.

## Conclusions and implications

The data examined above confirm that idea that KIBS are highly innovative industries. (The next paper in this series will consider how far they support innovation in their clients – and whether this can contribute to the transition towards more sustainable patterns of economic activity.) The analyses also support the view that it is T-KIBS that are outstanding in terms of innovation. KIBS’ innovation patterns vary across the different subsectors here. The evidence provides little support for the assumption that they concentrate more on organisational than on technological innovation.

The growth of KIBS has been a major source of new high-skill jobs in Western economies, and so far, there is little sign of this being seriously undermined by trends towards automation and offshoring. Often claims about the threat to employment of the sorts fond in KIBS are based upon alarmist projections that “at-risk” positions will be rapidly and thoroughly displaced. The prospects for clerical work may not be so positive; however, the long-term displacement of these jobs bears witness to considerable shifts in occupational structures even within expanding industries.

Claims that new generations of IT may be challenging more aspects of professional work deserve consideration. The trend data suggest that it is not just routine information-processing clerical jobs that have been impacted by earlier generations of IT; the balance between Professionals and Associate Professionals has also been shifting away from the latter. There is some evidence that many professional jobs are subject to greater management control, as well as involving use of tools that accomplish some of the work previously assigned to Associates. The content and conduct of professional work will almost certainly change, as:

• tools are developed to take on more sophisticated tasks; and

• networking and social innovation shifts expertise (for example, via social media) across organisations.

The quality of working life is liable to be affected, and the degrees of autonomy and creative problem-solving, and the fulfilment and commitment to work that is elicited, will be issues requiring attention and, hopefully, eliciting action.

In terms of social justice and political sentiment, perhaps the more immediate issue concerns polarisation trends. The analyses have indicated that such trends are evident within KIBS industries, and are not restricted to manufacturing and less knowledge-intensive services. What does the relative decline of intermediate positions mean for entry into KIBS professions, and for the diversity and inclusiveness of KIBS? One concern is that the route to entry into professional occupations via the gathering of experience in associate positions may be diminished, meaning that fewer opportunities would exist for those beginning with disadvantages to work their ways upwards. How far is this reflective of wider social trends – that may be contributing to the political divisiveness that has been characteristic of the second decade of this century – and what might it imply for the politics of expertise? Does this enhance the likelihood that professions are widely seen as exclusive and self-perpetuating clubs, filled by privileged people with little understanding of the heard realities faced by those whose employment prospects remain far more restricted? What does this imply for social trust and social cohesion?

More mundane issues of policy and strategy are also raised by these analyses. The trends and potential developments examined have implications for planning for:

• skill requirements (including the need for immigrants, as well as trained natives, as sources of skilled labour);

• education policies (for instance, training in creative problem-solving, early years education to offset social disadvantage, etc.);

• staff recruitment practices (especially those taking into account changing job content and the requirements of diverse and complex societies);

• professional training and broader staff support services (including opportunities for life-long learning);

• governance of data (and explicit examination of the ways in which machine learning and artificial intelligence applied to data sets may reproduce social inequalities); and

• establishment of professional standards and certification, along with improved ways in which users can assess quality of services and access and contribute to new forms of service provision.

The trends and prospects outlined above also offer opportunities for cultivation of human-centred innovation, development of new services for businesses and consumers and addressing work–life imbalances. These opportunities will require more subtle analysis of developments than is typically the case at present. It will be important to move beyond simply applying computerisation to displace jobs or the routine elements of jobs. We should not expect KIBS to move uniformly toward more standardisation or toward closer tailoring to client requirements. Beyond considerations of working life, the conclusions reached above need to be seen in the light of the contribution that KIBS may make towards the reorientation of innovation systems towards more sustainable trajectories. That issue will feature more prominently in the third of this series of essays.

## Figures

#### Figure 1

Share of firms in (a) Germany and (b) the UK undertaking major classes of innovation, 2014

#### Figure 2

Comparisons of shares of innovation introduced in various sectors (a) product (b) process (c) marketing and (d) organizational in Germany and the UK, 2014

#### Figure 3

Innovation specialisation, (a) Germany and (b) UK, 2014

#### Figure 4

Structure of expenditures on innovation across different KIBS sectors, UK and Germany, 2014

#### Figure 5

Occupational structures of three types of KIBS industry in 2017, UK

#### Figure 6

Occupations (ANZSCO classification) in professional, scientific and technical services (ANZSIC classification), Australia, 1986-2017

#### Figure 7

Ratios of support workers to professionals in professional, scientific and technical services, Australia 2008-2017

#### Figure 8

Ratios of support staff to professionals, EU15, (a) Germany and (b) UK

#### Figure 9

Detailed occupations (ISCO 2-4 categories) in the UK economy in 2017

#### Figure 10

Shares of workforce with different skills level in NACE section M and total economy for (a) EU-15, (b) Germany, (c) France and (d) UK (in %)

#### Figure 11

Changes of workforce with different skills level in 2017 compared to 2008 in NACE section M and total economy for EU-15, Germany, France and UK (in %)

#### Figure 12

Distribution of job types within UK industries, 2008-2017

#### Figure 13

Percentage increase in numbers of employees in job types, 2008-2017

## Table I

Extent to which KIBS specialize as suppliers of relevant knowledge, United States 2012 and 2017

KISA Occupation(s) Key KIBS Sector Share of Employmentin Key KIBS Sector
2012 2017Other sectors featuring > 10 per cent of the occupation(s) in 2017Share of sectors with <10 per cent of the occupation(s) in 2017
(F = 100-D-E)
Accountants, Auditors Accounting, Tax Preparation, Bookkeeping, and Payroll Services 27.31 26.04 None 73.96
Architects (Except Naval) Architectural, Engineering, and Related Services 81.64 83.20 None 16.80
Computer and Information Analysts Computer Systems Design and Related Services 28.29 30.25 Finance & Insurance (12.23) 56.51
Software Developers and Programmers Computer Systems Design and Related Services 33.54 34.07 Information* (17.66) 48.27
Designers Specialized Design Services 11.51 11.65 Retail (21.96)
Manufacturing (12.76) Wholesale (12.48)
41.15
Engineers Architectural, Engineering, and Related Services 20.79 21.19 Manufacturing (35.52)
Government (11.59)
31.70
Lawyers Legal Services 65.27 61.85 Government (21.38) 16.77
Marketing Research Analysts and Specialists + Public Relations Specialists** Advertising, Public Relations, and Related Services 8.61 8.12*** None 91.88
Management Analysts Management, Scientific, and Technical Consulting Services 25.85 24.63 Government (19.49)
Finance & insurance (12.45)
43.42
Medical Scientists + Biological Scientists Scientific Research and Development Services 30.68 30.95 Government (18.31)
Health Care and Social Assistance (16.62)
Educational Services (16.16)
17.96
Notes:

Figures are percentage of the workforce in the occupational group (column A), employed in the sector (column B). KIBS refers to the industry NAICS 54. *The Information sector includes Data processing, hosting, and related services, as well as Publishing and Broadcasting industries (except internet), Motion picture and sound recording, Telecommunications, and “other information services”. Parts of this could, arguably, also belong to the key KIBS sector for these Knowledge-Intensive Service Activities. ** There are actually more PR specialists than marketing specialists recorded in this sector, and more advertising sales agents than either of these groups, and more designers than marketing specialists. *** In 2017 another KIBS sector (Management, Scientific, and Technical Consulting Services) captures 9.74% of those occupations – more than what we presume to be the core KIBS specialist sector. In second place, also just above this core sector, was Finance and Insurance, with 8.14% of all Marketing Research Analysts & Specialists and Public Relations Specialists

Source: Occupational Employment Statistics (Bureau of Labour Statistics, USA) www.bls.gov/oes/tables.htm (accessed 31 October 2018)

## Notes

1.

However, Hipp et al. (2015) also include the Labour Recruitment industry within their set of KIBS, while this would generally be regarded more as an operational/administrative business service

2.

We have some reservations about this identification of product and process as technological, organisation and marketing as nontechnological. It is possible for there to be nontechnological product innovation (especially where the product is a service) and process innovations may involve restructuring of work processes; and of course, organisation and marketing innovation may involve new technology (e.g. digital media, knowledge management systems).

3.

It is generally considered that the UK labour market is typically much less regulated than its counterparts such as in Germany (e.g. Kananen, 2005), across Europe, which might enable more rapid or extensive work-related organisational change in response to economic events. While systematic comparative analyses of recent organisational developments – such as the rise of the “gig economy” – requires further development, it is possible to contrast some features of the UK and Germany through various indicators reported by Eurofound. For example, we see that data on the proportion of young people (of age 15-24) on non-standard contracts of employment in 2012 vary greatly across the countries: in Germany over 52 per cent are on temporary contracts – in the UK the figure is under 14 per cent. (Though few UK youth are “independent” workers, at 4.7 per cent this is over three times the share in Germany). When we look at the main reason for being in temporary employment, in Germany 84.3 per cent were undergoing a training period, as opposed to 10.4 per cent in the UK, where 47.4 per cent reported being unable to find permanent work and 36.4 per cent “did not want it” (presumably for the sort of job on offer)- compared to 6.7 and 1.5 per cent, respectively in Germany. Clearly, labour market dynamics are extremely different in the two countries (See Eurofound, 2013b). On zero hours’ contracts, rapidly increasing in the UK and very rare in Germany, see Eurofound (2015a).

4.

German data for 2016 became available while this essay was in preparation (UK data not yet published), and that indicates considerable volatility over even this short stretch of time. In general, innovation levels were higher, though some sectors decreased (particularly for organizational innovation); while the overall patterns look quite similar, there is far from perfect correlation in the order of the sectors across the two years. It is thus unwise to rely too much on the stability of sectoral propensities.

5.

This could be investigated by comparing CIS results from successive surveys in the two countries.

6.

We would like to draw attention to one study in particular, that of Love et al. (2011), which is one of the few survey studies to recognise that the innovation process extends over time, with different capabilities and resources being involved at different stages. They report that business services’ openness towards clients is crucial for innovation at the beginning stage of the project, while internal openness (teamwork within the firm) is needed at the later stages. See also Doroshenko (2012).

7.

Eurostat provides information that allows us to include French firms into some of this analysis. Data on expenditure levels is available for Germany, where the bulk of innovation expenditure is recorded as coming from the manufacturing sector, which is also higher in innovation expenditure as a share of turnover, or per employee, than other industries – except for M72, R&D services, which turn in very high figures, and J62, computer services (higher expenditure as share of turnover, lower per employee). J63 and M74 also report fairly high levels of expenditure weighted in these terms.

8.

On the concerns about automation, employment and the service economy, see Bowen and Mangum (1966), Bell (1973). On deskilling, the classic account was Braverman (1974; see the 1998 edition for an interesting introduction by Foster). Braverman’s analysis mainly concerned factory work, though he also saw a degradation of clerical work into simple data entry and similar trends in retail service and the “middle layers of employment”. We discuss literature on polarisation below, and consider Autor’s “task-based” approach in which occupations are seen as bundling together various tasks (e.g. Autor, 2013).

9.

In footnote 7. In summary, and rounded up, KIBS sectors (firms reporting innovation only) are reported for 2016 as spending, in thousands of euro per employee, as follows: M72- 47; J62 - 13; j63 - 6; M74 - 5; M73 - 3; M71 - 3; M70- 2; M69 - 1; as compared to section C - 18, section D - 16; section K - 6; section H - 5; section B - 4; and section E- 3.

10.

And for its member states, but there is not space to discuss the country variations here.

11.

The report presents some country-level data, but we here report figures for the EU28. There is no specific information on KIBS, but section J (information and communication) is seen as expanding at 0.8 per cent per annum, while sections L&M combined (i.e. real estate together with professional and scientific services) at a remarkable 1.6 per cent.

12.

The survey questions that they draw upon involve the features: codifiability; routineness; multitasking; complementarity, IT use; interactivity (e.g. with clients); locational ties; cultural ties (to do with knowledge of law, writing skills, language skills); having to deal with new problems. See Brändle and Koch (2014) for more detail, plus some information on trends in these job characteristics in the last decade – most striking in the case of writing and language skills.

13.

Some of these will be employed in KIBS, already effectively being outsourced.

14.

A recent study (Reijnders and de Vries, 2018) looks at the organization of work in global value chains, using input-output analyses. Their conclusion is that both technological change and (to a lesser extent) task relocation increased the number of non-routine relative to routine occupations in advanced countries like Germany and the United States but that task relocation works in the opposite direction for emerging countries like China and Poland.

15.

van Welsum and Reif (2009) provided analysis of data for exports and imports of “business and computer and information services”, in 1995 and 2003, for a number of OECD countries; this interesting study supplies more detail on differences across EU countries.

16.

Or that employers are using qualifications as a way of assessing the motivation of workers rather than their technical skills.

17.

Martinez-Fernandez et al. (2011) discuss these and other data in relation to KISA.

18.

A blog post from Eurofound contrasts data for 1995 and 2015, indicating the same decrease in routine jobs, but increase in routine job content: see https://www.eurofound.europa.eu/publications/blog/fewer-routine-jobs-but-more-routine-work (accessed 26 October 2018).

19.

The British SOC classifications follow ISCO categories rather closely. In what follows we will discuss numbers and shares of employed people, without delving into the rather important issue of whether these are full-time or part-time employees, or employed by others versus self-employed. There are significant variations across sectors and also across occupational categories; for example the British data show that around a tenth of employed managers are part-time workers, while a fifth of professionals are. Part-time work seems less common among the self-employed: around a quarter of self-employed managers are part-time, and a third of self-employed professionals.

20.

For this latter category we combine data from M73 (advertising) and M74 (other) KIBS subsectors.

21.

The ISCO job titles do not give us a comprehensive picture of what the job roles really entail. We can consider business firms as featuring three main sets of functions: (1) Production of their main product – in the case of KIBS, services for clients; (2) Maintenance of essential business processes – such as internal finances and routine management; (3) Strategic planning and similar senior-level activities. While we can see (1) as mainly involving professionals and associates, with some clerical support; (2) as mainly involving administrative and clerical staff; and (3) as mainly involving senior managers, this is only a rough approximation. There are professional and associate roles in functions (2) and (3) – for example, running corporate information systems. There are clerical roles in (1) and (3) – providing routine support for professionals and managers. In smaller firms, senior managers may be involved in (1), interfacing with clients; in the case of major clients, this may also apply in large KIBS. Thus, the discussion of ratios of various types of staff to professional in the following paragraphs should not elide the probability that many of the support services will not be directly supporting function (1) activities.

22.

This could be established by considering labour flows into and out of occupational roles.

23.

The dataset presented in Eurofound (2016, p. 48, Table 5) does allow us to compare the tasks undertaken by Associates, as compared to Professionals in various areas of expertise, for a single point in time (2015). We find, not surprisingly, that the former typically feature higher levels of physical skill, and lower levels of intellectual and social skills in their work; they also have less autonomy and more repetitive work, but not always more standardized work. A fuller report on these analyses will be published in due course.

24.

See Arundel et al. (2006) for interesting analyses of varieties of organisational culture across Europe.

25.

Goos et al. (2014) classify occupations into three groups based on wages using more detailed occupational categories data and excluding some categories (Legislators and senior officials; Teaching professionals; Teaching associate professionals; Skilled agricultural and fishery workers; Agricultural, fishery and related labourers). However, Eurostat provides only aggregated data. We adapted the classification of Goos et al. (2014), being in line with Eurostats’ data.

26.

Consoli and Roy (2015), working with data for regional employment districts, similarly note that the German pattern of development differs from the Anglo-Saxon one.

27.

The correspondence between 2008 and 2017 data is manually developed because of the changes in the Office for National Statistics’ classification of occupations.

28.

In the UK, the rate of decline of MS in section M is somewhat lower than in the whole economy; for the EU15, we see higher PS and LS expansion in section M than for the whole economy, with a slight growth in MS positions.

## References

Ahlqvist, T. (2003), Keys to Futures: Societal Reflections on Developing Key Technologies and Their Impacts on Human Qualifications, Ministry of Trade and Finance, Helsinki.

Arundel, A., Lorenz, E., Lundvall, B.-A. and Valeyre, A. (2006), “The organization of work and innovative performance: a comparison of the EU-15”, DRUID Working Paper No. 06-14, Copenhagen Business School, Department of Industrial Economics and Strategy/Aalborg University, Department of Business Studies, Copenhagen/Aalborg, Druid.

Asikainen, A.L. (2015), “Innovation modes and strategies in knowledge intensive business services”, Service Business, Vol. 9 No. 1, pp. 77-95.

Autor, D. (2013), “The task-approach to labor markets: an overview”, Journal for Labour Market Research, Vol. 46 No. 3, pp. 185-199.

Baldwin, J.R. (1995), “Innovation: the key to success in small firms”, Statistics Canada Working Paper 76, Statistics Canada, Ottawa, Ontario.

Bell, D. (1973), The Coming of Post-Industrial Society. A Venture in Social Forecasting, Heinemann, London.

Berger, T. and Frey, C. (2016), “Structural transformation in the OECD: digitalisation, deindustrialisation and the future of work”, OECD Social, Employment and Migration Working Papers No. 193, OECD Publishing, Paris.

Bolisani, E., Paiola, M. and Scarso, E. (2014), “Cognitive features of KIBS companies: evidence from a cluster analysis”, International Journal of Services Technology and Management, Vol. 20 Nos 4/5/6, pp. 215-232.

Bowen, H. R. and Mangum, G. L. (Eds) (1966), Automation and Economic Progress, Prentice-Hall, Englewood Cliffs, NJ.

Braga, A.M. and Marques, C.S. (2016), “Knowledge intensive business services research: a bibliometric study of leading international journal (1994-2014)”, in Ferreira, J., Raposo, M., Fernandes, C. and Dejardin M. (Eds), Knowledge Intensive Services and Regional Competitiveness, Routledge, New York, NY, pp. 11-47.

Brändle, T. (2015), “Is offshoring linked to offshoring potentials? Evidence from German linked-employer-employee data”, IAW-Diskussionspapiere, No. 112, Institut für Angewandte Wirtschaftsforschung (IAW), Tübingen.

Brändle, T. and Koch, A. (2014), “Offshoring and outsourcing potentials of jobs: evidence from German micro-level data”, IAW-Diskussionspapiere, No. 110, Institut für Angewandte Wirtschaftsforschung (IAW), Tübingen.

Braverman, H. (1998), Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century, 25th anniversary edition, Monthly Review Press, New York, NY.

Cedefop (2016), “Future skill needs in Europe: critical labour force trends”, Cedefop research paper No 59, Publications Office of the European Union, Luxembourg.

Cefis, E. and Marsili, O. (2005), “A matter of life and death: innovation and firm survival”, Industrial and Corporate Change, Vol. 14 No. 6, pp. 1167-1192.

Consoli, D. and Elche, D. (2013), “The evolving knowledge base of professional service sectors”, Journal of Evolutionary Economics, Vol. 23 No. 2, pp. 477 -501.

Consoli, D. and Roy, D. (2015), “Employment polarization in Germany: role of technology, trade and human Capital”, Jena Economic Research Papers No. 2015-017, Friedrich Schiller University Jena, Jena.

Doloreux, D. and Shearmur, R. (2013), “Innovation strategies: are knowledge-intensive business services just another source of information?”, Industry and Innovation, Vol. 20 No. 8, pp. 719-738.

Doroshenko, M.E. (2012), “How knowledge-intensive business services upgrade their customers: evidence from Russia”, in Di Maria, E., Grandinetti, R. and Di Bernardo, B. (Eds), Exploring Knowledge-Intensive Business Services, Palgrave Macmillan, London, pp. 79-99.

EMCC (2006a), “Trends and drivers of change in the knowledge-intensive business services sector: four scenarios”, (accessed 31 October 2018).

EMCC (2006b), “Trends and drivers of change in the European knowledge-intensive business services sector: mapping report”, (accessed 31 October 2018).

Eurofound (2013a), Employment Polarisation and Job Quality in the Crisis: European Jobs Monitor 2013, Eurofound, Dublin.

Eurofound (2013b), “Young people and temporary employment in Europe”, (accessed 31 October 2018).

Eurofound (2015a), New Forms of Employment, Publications Office of the European Union, Luxembourg.

Eurofound (2015b), Upgrading or Polarisation? Long-term and Global Shifts in the Employment Structure: European Jobs Monitor 2015, Publications Office of the European Union, Luxembourg.

Eurofound (2016), What Do Europeans Do at Work? A Task-based Analysis: European Jobs Monitor 2016, Publications Office of the European Union, Luxembourg.

Felstead, A., Gallie, D., Green, F. and Inanc, H. (2013), “Work intensification in Britain: first findings from the skills and employment survey 2012”, (accessed 31 October 2018).

Felstead, A., Gallie, D., Green, F. and Ying, Z. (2007), “Skills at work, 1986 to 2006”, (accessed 31 October 2018).

Frey, C.B. and Osborne, M.A. (2013), “The future of employment: how susceptible are jobs to computerisation?”, (accessed 31 October 2018).

Frey, C.B. and Osborne, M.A. (2017), “The future of employment: how susceptible are jobs to computerisation?”, Technological Forecasting and Social Change, Vol. 114, pp. 254-280.

Goos, M., Manning, A. and Salomons, A. (2014), “Explaining job polarization: routine-biased technological change and offshoring”, American Economic Review, Vol. 104 No. 8, pp. 2509-2526.

Gotsch, M., Hipp, C., Gallego, J. and Rubalcaba, L. (2011), “Sectoral innovation watch: knowledge intensive services sector”, (accessed 31 October 2018).

Henseke, G., Feldstead, A., Gallie, D. and Green, F. (2018), “Skills trends at work in Britain: first findings from the skills and employment survey 2017”, (accessed 31 October 2018).

Hipp, C. and Grupp, H. (2005), “Innovation in the service sector: the demand for service-specific innovation measurement concepts and typologies”, Research Policy, Vol. 34 No. 4, pp. 517-535.

Hipp, C., Gallego, J. and Rubalcaba, L. (2015), “Shaping innovation in European knowledge-intensive business services”, Service Business, Vol. 9 No. 1, pp. 41-55.

Howells, J., Tether, B., Gallouj, F., Djellal, F., Gallouj, C., Blind, K., Edler, J., Hipp, C., Montobbio, F., Corrocher, N., Macpherson, A. and Banach, D. (2004), “Innovation in services: issues at stake and trends”, (accessed 31 October 2018).

Inanc, H., Felstead, A., Gallie, D. and Green, F. (2013), “Job control in Britain: first findings from the skills and employment survey 2012”, (accessed 31 October 2018).

J-Figueiredo, R., Neto, J.V., Quelhas, O.L.G. and Ferreira, J.J.M. (2017), “Knowledge intensive business services (KIBS): bibliometric analysis and their different behaviors in the scientific literature”, RAI Revista de Administração e Inovação, Vol. 14 No. 3, pp. 216-222.

Kananen, J. (2005), “Current employment policy paradigms in the UK, Sweden and Germany”, in Taylor-Gooby, P. (Ed.), Ideas and Welfare State Reform in Western Europe, Palgrave Macmillan, London, pp. 124-144.

Liu, R. and Trefler, D. (2008), “Much ado about nothing: American jobs and the rise of service outsourcing to China and India”, NBER Working Paper Series 14061, National Bureau of Economic Research, Cambridge, MA.

Love, J.H., Roper, S. and Bryson, J.R. (2011), “Openness, knowledge, innovation and growth in UK business services”, Research Policy, Vol. 40 No. 10, pp. 1438-1452.

Mansury, M.A. and Love, J.H. (2008), “Innovation, productivity and growth in US business services: a firm-level analysis”, Technovation, Vol. 28 Nos 1/2, pp. 52-62.

Martinez-Fernandez, C., Miles, I. and Weyman T. (Eds) (2011), The Knowledge Economy at Work: Skills and Innovation in Knowledge Intensive Service Activities, Edward Elgar, Cheltenham; Northampton, MA.

Miles, I. (2005), “Knowledge intensive business services: prospects and policies”, Foresight, Vol. 7 No. 6, pp. 39-63.

Miles, I. (2008), “Patterns of innovation in service industries”, IBM Systems Journal, Vol. 47 No. 1, pp. 115-128.

Miles, I.D., Belousova, V. and Chichkanov, N. (2018), “Knowledge intensive business services: ambiguities and continuities”, Foresight, Vol. 20 No. 1, pp. 1-26.

Moncarz, R.J., Wolf, M.G. and Wright, B. (2008), “Service-providing occupations, offshoring, and the labor market”, Monthly Labor Review, Vol. 12, pp. 71-86.

Nählinder, J. (2005), “Innovation and Employment in Services: The Case of Knowledge Intensive Business Services in Sweden”, Doctoral thesis, Department of Technology and Social Change, Linkoping University, (accessed 31 October 2018).

O*NET (2006), New and Emerging (N&E) Occupations: Methodology Development Report, available at: www.onetcenter.org/dl_files/NewEmerging.pdf (accessed 31 October 2018).

Reijnders, L.S.M. and de Vries, G.J. (2018), “Technology, offshoring and the rise of non-routine jobs”, Journal of Development Economics, Vol. 135, pp. 412-432.

Rodgers, R. and Waters, R. (2001), “The skills dynamics of business and public service associate professionals”, available at: http://dera.ioe.ac.uk/4587/1/RR302.PDF (accessed 31 October 2018).

Rounds, J., Su, R., Lewis, P. and Rivkin, D. (2013), “Occupational interest profiles for new and emerging occupations in the O*NET system: summary”, available at: www.onetcenter.org/dl_files/OIP_NewEmerging.pdf (accessed 31 October 2018).

Sako, M. (2014), “Outsourcing and offshoring of professional services”, in Empson, L., Muzio, D., Broschak, J.P. and Hinings, B. (Eds), The Oxford Handbook of Professional Service Firms, Oxford University Press, Oxford.

Scarso, E. (2015), “What do we know about KIBS? Results of a systematic literature review”, in Spender J., Schiuma G. and Albino V. (Eds), Proceedings of IFKAD 2015, pp. 1159-1172, (accessed 31 October 2018).

Schmidt, T. and Rammer, C. (2007), “Non-technological and technological innovation: strange bedfellows?”, ZEW Discussion Paper No. 07-052, ZEW Centre for European Economic Research, Mannheim.

Schricke, E., Zenker, A. and Stahlecker, T. (2012), Knowledge-intensive (business) services in Europe, Publications Office of the European Union, Luxembourg.

Shearmur, R. and Doloreux, D. (2013), “Innovation and knowledge-intensive business service: the contribution of knowledge-intensive business service to innovation in manufacturing establishments”, Economics of Innovation and New Technology, Vol. 22 No. 8, pp. 751-774.

Sundbo, J. and Gallouj, F. (2000), “Innovation as a loosely coupled system in services”, in Metcalfe, S. and Miles, I. (Eds), Innovation Systems in the Service Economy: Measurement and Case Study Analysis, KIuwer Academic Publisher, New York, NY, pp. 43-68.

Susskind, R. and Susskind, D. (2015), The Future of the Professions: How Technology Will Transform the Work of Human Experts, Oxford University Press, Oxford.

Tödtling, F., Lehner, P. and Kaufmann, A. (2009), “Do different types of innovation rely on specific kinds of knowledge interactions?”, Technovation, Vol. 29 No. 1, pp. 59-71.

Van Welsum, D. and Reif, X. (2009), “We can work it out: the globalisation of ICT-enabled services”, in Reinsdorf, M. and Slaughter, M.J. (Eds), International Trade in Services and Intangibles in the Era of Globalisation, University of Chicago Press, Chicago, IL, pp. 289-325.

Winterton, J., Parker, M., Dodd, M., McCraken, M. and Henderson, I. (2000), “Future skills need of managers”, available at: www.dfee.gov.uk/research/index (accessed 31 October 2018).

## Acknowledgements

This paper was prepared based on research carried out as part of the Basic Research Programme of the National Research University Higher School of Economics (NRU HSE) and was supported by a state subsidy granted to the NRU HSE to help leading universities in the Russian Federation become more competitive global scientific research centres.

## Corresponding author

Ian Douglas Miles can be contacted at: IMiles@HSE.ru