A question of context

In international comparisons of the CIS 2008 data, Norway ranks relatively low among the countries conducting the survey. In the Innovation Union Scoreboard1 report published by the EU DG Enterprise, Norway is ranked below the EU average and placed in the third out of four groups as a "moderate innovator". Comparatively, the other Nordic Countries, Denmark, Sweden, and Finland are all belonging to the primary category of "leading innovator". Norway is also categorized as a "slow growing innovator", as is Denmark end Sweden.

Yet, Norway does very well when it comes to core economic indicators such as comparatively high GDP per capita growth, a high overall trade balance surplus, low unemployment etc. If we expect a direct causal relationship between a high score on innovation indicators and economic prosperity, this may seem like a paradox. Thus, it seems pertinent to question the validity of benchmarking on indicators where Norway scores particularly poorly, yet unequivocally does very well on the overall economic statistics that these are supposed to be input to - or explanatory factors for.

Other reports have discussed this issue several times, perhaps chiefly the OECD (2007), terming the phenomenon "the Norwegian Puzzle". While part of the Norwegian puzzle is indeed understood, it would still seem that some of the Norwegian results are lower than they reasonably "should be" when comparing the Norwegian economy to other countries. This is particularly true for some CISderived indicators.

This work presents a project seeking to explore whether there exists methodological factors that can help further enlighten the Norwegian puzzle. In particular, we wanted to answer two separate - but nevertheless related - questions.

First: Can the presence of detailed R&D questions influence the reported incidence of innovation? Most countries have separate R&D and innovation surveys, while some - including Norway - have integrated them in a single combined survey. The concept and measurement of innovation has been evolving rapidly, away from what was initially a purely technology and R&D driven focus, so there is a possibility that a combined survey may limit the respondents' understanding of what constitutes an innovative activity.

Second: Can we identify an effect of having an enforced mandatory innovation survey? Norway has traditionally had among the highest response rates among the countries carrying out the CIS, and there are differences when it comes to the survey conductors whether they employ a mandatory vs. a voluntary survey. While accurate in theory, there is cause for doubt about the comparability of results with and without non-response adjustment (as well as various ways of performing nonresponse surveys), and other countries have seen a drop in their innovation rates after making the CIS mandatory.

The results show that both these concerns are valid. We find a significantly higher share of innovators in a special sample having received a survey questionnaire covering only innovations and not R&D as compared to the results from a corresponding sample from the regular, combined R&D and innovation survey. Moreover, we find that the reported innovation rates increase even further when looking at an additional sample where the same innovation-only survey was also made voluntary. In total, the measured incidence of product and/or process innovation more than doubled going from a mandatory combined R&D and innovation survey to a voluntary innovation survey alone.