5.3. Innovation efficiency: linking inputs to outputs

This section provides a summary of the thematic paper on innovation efficiency [1]. Following the Lisbon strategy and the Barcelona target of an R&D intensity of 3% in 2010, many countries have taken steps to increase their innovation efforts. Innovation efficiency is related to the concept of productivity. Higher productivity is achieved when more outputs are produced with the same amount of inputs or when the same output is produced with less input. Innovation efficiency will here be defined similarly: innovation efficiency is improved when with the same amount of innovation inputs more innovation outputs are generated or when less innovation inputs are needed for the same amount of innovation outputs. Although innovation is not a linear process where inputs automatically transfer into outputs, it is worthwhile to examine differences in efficiency by assuming that efficiency can be defined as the ratio of outputs over inputs. In the EIS the indicators are divided into 3 innovation input dimensions covering 15 indicators and 2 innovation output dimensions covering 10 indicators [2]. Innovation efficiency will be measured by comparing the ratio between the composite indicator scores for one or more input dimensions and one or more output dimensions. Inputs and outputs can be plotted in a multidimensional space where the most efficient performers will be on or close to the ‘efficiency frontier’. The larger the distance to this frontier, the smaller will be the level of innovation efficiency [3]. In a two-dimensional graph with inputs on one axis and outputs on the other axis, the frontier can be visualised as the envelope curve connecting those dots with the most efficient output/input ratios.

In our analysis we have employed a constant-returns-to-scale output-oriented DEA (Data Envelopment Analysis [4]) on all combinations of the 3 input and 2 output dimensions. Missing values have been imputed using the techniques used in the 2005 EIS Methodology Report. The analyses were done separately for the most innovative countries (Sweden, the innovation leaders and innovation followers) and for the least innovative countries (moderate innovators and catching-up countries). Average efficiency scores for both output dimensions are shown in Figure 10.

All innovation leaders except Sweden have above average efficiency in transforming inputs into Applications. Despite its overall leadership in innovation performance, Sweden has the lowest efficiency in Applications of these countries indicating that despite its very gpood overall performance it has room to make improvements here. Germany and Switzerland show high efficiency in generating Intellectual property. Some of the innovation leaders, in particular the UK, have relatively low efficiency in transforming inputs into Intellectual property outputs. This may because the type of their innovation activities does not lead to formal IPRs but it could also indicate that these countries could be creating more IPRs for their level of inputs.
The innovation followers have above average efficiency in transforming inputs into Applications, with Luxembourg and Belgium showing highest efficiency rates. Only Austria, the Netherlands and Luxembourg show above average efficiency in Intellectual property, and hence Belgium, France and Iceland could seek to improve their efficiency rates by generating more IPRs from their innovation inputs.

The moderate innovators show a range of different efficiencies: we find these countries in all four quadrants in Figure 10 combining above or below average efficiency performance. Italy combines above average efficiency scores in both output dimensions. This result suggests that it may be difficult for Italy to improve its innovation performance without increasing innovation inputs. Australia, Cyprus, Norway and Spain show above average efficiency in Intellectual property [5] and the Czech Republic shows above average efficiency in Applications. Estonia and Slovenia combine below average efficiency in both Applications and Intellectual property.

The catching-up countries also show a variety of efficiencies in transforming innovation inputs into Applications. On Intellectual property efficiency all countries are significantly below average with the exception of Portugal. This may be because IPR is of less relevance for the innovative activities of these countries or that there is the potential to generate higher levels of IPR from the existing inputs. Some of these countries are also still in a process of replacing national patent applications by EPO patent applications. For Slovakia and Romania the efficiencies for Applications are relatively high, suggesting that these countries need to increase inputs to increase performance in generating more Applications. The majority of catching up countries have below average efficiencies and this suggests that for these countries an important focus should be improving innovation efficiencies.

Based on their relative position in Figure 10, peer countries in efficiency terms can be identified as those countries with higher efficiency scores in either Applications or Intellectual property. For example, Austria's possible peer countries include Germany, Luxembourg, the Netherlands and Switzerland, which combine higher efficiency scores in both Applications and Intellectual property. The innovation policies implemented in these countries could be compared with those in Austria to identify options for policy improvements to improve the efficiency of transferring innovation inputs into outputs. [6]