Understanding the Economics of Sport through the Industry Space Methodology

 

The Center for International Development at Harvard University in cooperation with Anet Weterings from the PBL Netherlands Environmental Assessment Agency is preparing a publication which investigates the sports cluster in the industry space. This blog highlights some of the findings of this study. This research is part of an ongoing collaboration with the International Centre for Sport Security (ICSS).

Any casual observer of the Olympics, the World Cup, or European professional soccer knows that the world of sport is big business. Expensive stadiums and massive player transfer fees suggest that there is an economic dimension to sport. However, the exact mechanisms through which sport influences other sectors in the broader economy are less clear. The industry space is one technique through which we can better understand these connections. It is a methodology that illustrates the industrial structure of a given geographic area by emphasizing the linkages between economic activities. Economic sectors are connected in the industry space if they possess similar characteristics such as human capital, natural resources, or value chain linkages. If we can locate sport within the industry space, we can hope to better understand how the economic aspects of sport spillover to other sectors.

Figure 1: Sports sectors in the industry space (Click all to enlarge)

Source: Data from statistics Netherlands, edited by PBL. Own calculations of the industry space.

To illustrate the industry space technique, we used data from the Netherlands from 2001-2008 on the number of people employed in 826 different economic activities and the number of employees switching jobs between these activities. We utilize Dutch data due to its greater level of detail in the classification of sports-related activities. Each of these economic activities is represented by a node depicted in the industry space in Figure 1. The network in Figure 1 is constructed based on labor flows between economic sectors such that the links between the nodes demonstrate the skill relatedness between these economic activities. The full network has many more links amongst the nodes, but, for visualization purposes, we only show the strongest 2,478 linkages, which is three times the number of the nodes in the network. The size of the node is proportional to the employment in that sector and the color of the node represents a cluster or community of economic sectors. These communities are collections of sectors with a high level of mutual labor flows. They are constructed using community detection algorithms from social network science (here we use the Fast & Greedy algorithm). Workers are more likely to move to sectors within the same community because these industries require similar skills. It is comparatively more difficult for a worker to jump to a sector outside of his or her community because that sector likely requires skills that the worker doesn’t already possess.

Of the 826 activities discernible in the data, 28 activities pertain to sport in the industrial classification that we use (SBI93, revision 2003). In Figure 1, sports-related activities have a triangular shape to distinguish them from the rest of the economic activities. The sports activities can be placed in 4 categories. First, 23 of the activities pertain to sports clubs such as soccer, tennis, or track and field. Second, 5 activities relate to sports facilities such as swimming pools, gyms, or sports halls. Third, 3 activities pertain to support organizations like fan associations. The final category includes professional sportsmen and sports instructors. The industry space depicted in Figure 1 reveals a significant amount of information about the role of these sports-related activities in the broader Dutch economy.

Figure 2: Sports clusters

To start, the industry space demonstrates how sports co-occur or group with other industries. In particular, two communities stand out when it comes to sports. First, 21 of the 28 sports-related activities cluster in a single community of red nodes depicted in Figure 2, confirming that, relative to all other economic activities, sports share a very similar skills base. Sports-related activities in the community, which are represented by the triangular nodes, include soccer, swimming pools, and motor sports. Besides these activities, the community primarily consists of health care, social care, and educational services such as hospitals, child care, and universities. 

The purple nodes designate a second community that contains 4 other sports-related activities: winter sports, billiards, board games and puzzles, and other outdoor sports. These activities cluster with transportation, recreation, and accommodation activities like restaurants, cafeterias, and railways. The remaining 3 sports sectors, which do not belong to either of these two communities, belong to other clusters. For instance, equestrian activity is more closely linked to agriculture and farming than to other sports and is therefore part of the agricultural products community.

Figure 3: Swimming - an example of sports sector that mainly connects to its own cluster

In addition to these clusters, the industry space also reveals interesting information about the different types of connections that sports-related activities have. Many sports-related activities are poorly connected to the broader economy and link to very few other nodes. Cycling, for instance, connects to only two other economic sectors once we selected for the top 5 percent of the strongest links from the full network. At the same level of link selection, some sports-related activities are connected to many other nodes, but the sectors to which they link are mostly only other sports. Swimming is an excellent example. It has linkages to swimming pools, sports halls, martial arts, and other advisory sport groups (Figure 3). On the other hand, other sports activities may not be densely connected, but the connections that they do have are diverse and link them to several different types of communities. Motorsports, for example, relates to four different communities and has connections with a diverse array of nodes that include car repair, hotel restaurants, and physical well-being activities (Figure 4).

Figure 4: Motorsports - an example of sports sector with cross-cluster connections

Soccer is an example of a sports-related activity with linkages that are both numerous and diverse. Figure 5 depicts a visualization of soccer’s connections in the industry space. The figure zooms in on soccer’s node and shows the links it shares with other economic activities. The linkages are grouped by community and, within each community, the strength of the linkages decreases as one proceeds clockwise around the circle. The strongest connections are shown at the top of the ring. Unsurprisingly, employees in the soccer sector often work in other sports-related activities such as professional sportsmen, sports instructors, and sports facilities. However, the ring also shows that employees in the soccer sector work in numerous other communities. These communities include sectors such as newspaper publishing, textile sales, and building cleaning. This demonstrates that relatedness often cuts across clusters. While sports-related activities like soccer may be primarily linked to other sports, they also share many connections with the broader economy.

Figure 5: Soccer - an example of densely interlinked sports sector with connections to many clusters

Analyzing the location of sport in the industry space is important because it helps us understand how sport affects other economic sectors. When a sports sector experiences an economic shock such as a mega-event like the World Cup, the industry space can help predict which related sectors will be influenced. For instance, increased labor demand in a sports industry may result in labor poaching and increased wages in a connected sector. Which nodes and which communities are affected by these spillovers? How strong must the connections between sectors be for these spillover effects to occur? Alternatively, if a government wants to create a sports industry, the industry space can help identify what related industries are helpful for the establishment of a robust sports sector. Are medical activities helpful complements of a strong sports sector? If not, are education activities more useful? As our research on sport and economic complexity progresses, we hope to address many of these research questions.

About the authors: Ljubica Nedelkoska is a Research Fellow with CID's Growth Lab and Stuart Russell is a Program Assistant at CID.