1. Knowledge-intensive activities in and around Brussels

1What is the extent of knowledge activities in Brussels? Are they atypical when compared to the rest of Belgium? As simple as those questions may seem, knowledge-intensive activities can be defined in different ways [Galindo-Rueda and Verger, 2016; Marsili and Verspagen, 2002].

2Official statistics of the OECD and Eurostat rely on different sector classifications to define “knowledge industries”. Rather than working with a predefined list, we can assume that every sector hosts knowledge-intensive jobs and carry out an empirical count. This more detailed empirical approach captures knowledge intensity within a sector by combining both the educational attainment and the occupational status of employed persons. By combining labour force survey data with social security statistics, we consider a person who has at least a short-cycle higher education degree and who has an International Standard Classification of Occupations (ISCO) status of either “manager”, “professional” or “technician or associate professional” (i.e. ISCO levels 1, 2 and 3) as a knowledge-intensive worker within a sector [see Eyckmans et al., 2022].

3Based on the detailed breakdown described above, we obtain an average number of 323 677 knowledge-intensive jobs in the Brussels-Capital Region (BCR), in the 2008-2020 period. This implies that, in the year 2020, 45,2 % of the jobs in Brussels can be regarded as knowledge-intensive. In comparative terms, this makes the BCR the most knowledge-intensive region in Belgium. Similar intensities are observed around Brussels, in the wider Brussels Metropolitan Area (BMA – in light blue in figure 1), but the numbers for the rest of the Flemish (31,3 %) and Walloon (30,9 %) Regions are considerably lower.

4However, over the 2008-2020 period, compared to an average of 1,11 % at national level, the compound annual growth rate of knowledge-intensive employment was lowest for the BCR (0,51 %) and highest in the surrounding BMA (excluding Brussels). In order to better explain and understand the rates in Brussels, a shift-and-share analysis was carried out. This involves applying the growth rates of each sector in a reference area – in this case, defined as the rest of Belgium (Belgium without Brussels) – to the different territories (in this case, Brussels). The aim of the analysis is therefore to distinguish between a structural effect – which assumes that growth is the result of the sectoral composition of local employment – and a regional effect. The regional effect is measured by the gap between the theoretical growth in employment (deduced from sectoral composition) and the observed growth [Murray, 2010]. In the case of Brussels, sectoral composition is characterised by the strong presence of dynamic sectors on average, but these sectors perform less well in Brussels. The analysis for Brussels therefore indicates that, while several important knowledge-intensive sectors such as financial services, human health, professional, scientific, legal and accounting activities and activities of head offices are still overrepresented in Brussels, it is precisely in those sectors that Brussels lost ground in terms of employment growth relative to the rest of Belgium. The picture is a bit different with respect to added value, since sectors such as finance have experienced strong economic growth despite a significant decrease in the number of jobs in the 2010s.

2. Spatial distribution patterns of knowledge-intensive activities

5A municipality-based mapping presents the geographical distribution of highly knowledge-intensive activities in Belgium. It illustrates the concentration of such activities in and around Brussels, far ahead of Antwerp, the second biggest Belgian city in terms of knowledge activities, as well as the high level of specialisation of Brussels in such activities.

6This spatial distribution pattern can be further detailed according to specific sectors. It confirms the high concentration of several knowledge services in the BMA, with its large and qualified labour pool, attractiveness for international workers, connectivity at European and global level, airport and high-speed rail connections, etc.

7However, spatial analyses also indicate that the geographical concentration of knowledge-intensive activities in Belgium is sector specific. Concentration in the BMA ranges from around half of all Belgian jobs in finance to a third in other sectors, such as legal and accounting services. Additionally, within the BMA there are very different locational patterns: some clusters of activities (e.g. finance) are mainly located in the city centre, while others may favour a location in the outskirts (e.g. the pharmaceutical industry, with large plants in Walloon Brabant, or large accounting firms located in Zaventem). Thus, our analysis underscores that – just as with other economic activities – the “Brussels” knowledge base implies an unequal geographical spread within Brussels, and necessarily implies that one should also take the relationship with the outskirts of Brussels into account.

3. Knowledge-intensive activities in Brussels and the creation of high- and low- skilled jobs

8The impact of knowledge-intensive activities can be assessed through the lens of econometric models which examine the relationship between the growth of knowledge-intensive employment and total employment, i.e. a so-called (relative) employment multiplier [see for example Moretti, 2010, or Dotti et al., 2021].

9The estimated coefficient in a simple regression comes conceptually closest to the conventional definition of a multiplier. We find that each additional knowledge-intensive job in Brussels can be associated with the creation of another 2,12 jobs in the BCR. This is a relatively low figure when compared to other parts of Belgium (e.g. 2,44 jobs in the Flemish Region). However, it is important here to recall the relative nature of such multipliers: a lower ratio of total employment growth to total knowledge-intensive employment growth might as well be the result of a lower value for the numerator as the result of a comparatively high value of the denominator (and vice versa). And in fact, Brussels fares comparatively well, precisely in terms of the denominator (see above).

10In reality, the change in knowledge-intensive jobs is not exogenous, and several variables have an impact on total employment change. By taking into account sector composition and economic and knowledge-related indicators, the coefficient associated with knowledge-intensive jobs remains statistically significant: all else being equal, in the BCR an extra job in a knowledge-intensive activity generates an additional job in total employment.

11Employment multipliers can alternatively be derived directly from the interregional input-output tables of the Federal Planning Bureau. Consistent with the underlying economic (Leontief) model from which they are derived, these multipliers can be used to assess the employment effects following a demand shock which affects one or more knowledge-intensive sectors located in the BCR [Avonds et al., 2021].

12A virtue of the input-output framework is that in addition to breaking down direct and indirect employment effects, we are able to see the intra- and interregional (“spillover”) employment effects of such shocks. Moreover, based on a method discussed by Miller and Blair [2022, pp. 25-26], our empirical information on the knowledge intensity of sectors (see section 1) can be used to break down all of these effects further in the creation of knowledge-intensive jobs as well as less qualified jobs.

14Next, consider a 1 % demand shock affecting all knowledge-intensive sectors. This would lead to about 3000 extra jobs (the exact number depends on the definition of the relevant sectors discussed in section 1). Table 1 shows more details for the case in which the OECD definition is used to identify the sectors experiencing a demand shock.

Table 1. Aggregate employment impact of a 1 % increase in final demand of all knowledge-intensive sectors in Brussels (using the OECD definition)

Extra jobs:	Knowledge-intensive	Non-knowledge-intensive	Total
In BCR:	1618	1092	2710
Initial effect	1406	860	2266
Indirect effects	212	232	444
In the Flemish Region	118	248	366
In the Walloon Region	59	126	185
TOTAL Belgian Economy	1795	1466	3262