The amount and quality of knowledge circulating and spilling over between firms located in a cluster is dependent upon the cluster's size, the degree to which it is specialised and the extent to which the locality (the region) is geared towards and focused upon production in the relevant industries comprising the cluster. These three factors, size, specialisation and focus, reflect whether the cluster has reached 'specialised critical mass' to develop positive spill-overs and linkages. The European Cluster Observatory shows the extent to which clusters have achieved this specialised critical mass by employing measures of these three factors as described below, and assigning each cluster 0, 1, 2 or 3 'stars' depending on how many of the below criteria are met. The 32 countries in the analysis are referred to as 'Europe' for simplicity's sake; the countries comprise the EU-27 and the five associate countries of Iceland, Israel, Norway, Switzerland and Turkey.

- Size: if employment reaches a sufficient share of total European employment, it is more likely that meaningful economic effects of clusters will be present. The 'size' measure shows whether a cluster is in the top 10% of all clusters in Europe within the same cluster category in terms of the number of employees. Those in the top 10% will receive one star.

- Specialisation: if a region is more specialised in a specific cluster category than the overall economy across all regions, this is likely to be an indication that the economic effects of the regional cluster have been strong enough to attract related economic activity from other regions to this location, and that spill-overs and linkages will be stronger. The 'specialisation' measure compares the proportion of employment in a cluster category in a region over the total employment in the same region, to the proportion of total European employment in that cluster category over total European employment (see equation). If a cluster category in a region has a specialisation quotient of 2 or more it receives a star.



- Focus: if a cluster accounts for a larger share of a region's overall employment, it is more likely that spill-over effects and linkages will actually occur instead of being drowned in the economic interaction of other parts of the regional economy. The 'focus' measure shows the extent to which the regional economy is focused upon the industries comprising the cluster category. This measure relates employment in the cluster to total employment in the region. The top 10% of clusters which account for the largest proportion of their region's total employment receive a star.

If the number of employees in a cluster is less than 1,000 persons, the cluster is not given any stars to prevent the appearance of very small insignificant clusters.

Alternative approaches used in the literature are, for example, the measures of employment concentration (Gini coefficient or similar measures) or the share of employment in regional clusters identified as strong. The employment concentration measure can be applied either within the regional economy or within the cluster category across regions; in the former instance it comes close to our measure of 'Focus', in the latter, to our measure of 'Specialisation'. The measure for the share of employment in strong clusters comes close to a combination of our measures for 'Size' and 'Specialisation'. In our view the '3-star' approach offers a new way to combine these perspectives. The three approaches do give comparable results, although at a more detailed level some differences can occur.

Data limitations restrict us to the use of employment data to identify and evaluate clusters. This creates a certain bias in our measures towards employment-intensive clusters, especially on the metrics for size and focus. Only the measure for specialisation is unaffected by differences of employment intensity across cluster categories. It would have been preferable to use data on wage bill, productivity, or value added, which would have shifted the balance in favour of capital- or knowledge-intensive cluster categories such as biopharmaceuticals, but as such data was not available in all European countries we have resorted to the use of employment data.