The Role of the EU Internal Market in the Adoption of Innovation

Abstract

This paper aims at analysing the impact of the regulation of the European Union (EU) internal market (IM) in the adoption of innovation. After constructing an indicator of innovation adoption based on the information provided by the Community Innovation Survey, in the first stage, we define the impact of some major IM regulations on cooperation, competition and trade across EU countries. The results of this first stage show how different IM regulations are important determinants of these three macroeconomic variables that we consider afterwards having an impact on innovation adoption. Hence, in the second stage we address whether innovation adoption rates significantly depend on the degree of cooperation, trade and competition as well as some control variables such as national legal structures and IPR regulations. Estimations have been computed using an econometric model whose dependent variable is total innovation adoption as well as its possible disaggregation into sub-categories. Results show that cooperation and its main IM determinants positively impact innovation adoption, leaving trade and competition as apparently minor channels of innovation adoption (and especially depending on the type of innovation adoption under examination).

Acknowledgements

The authors would like to acknowledge the support of DG ECFIN/E/2007/020, CICYTs ECO2008-05314 and ECO2009-12678. Additionally, we are indebted to the comments by Corinne Autand-Bernard, Nadine Massard and Jacques Mairesse. The Third Community Innovation Survey, microdata, have been provided by the Eurostat, European Commission. Eurostat has no responsibility for the results and conclusions obtained in this paper. The views expressed in this paper are of the authors only and do not represent the official position of any of the affiliated institutions.

Notes

In some cases, it has been necessary to homogenize the information available from both micro and macro databases provided by the CIS.

That is, given the same amount of accumulated stock of technology in two different countries, the intensity of the returns stemming from it will be the same across countries. On this see the discussion of Jones (2002) p.99.

See for example the Institutional indexes proposed by Hall and Jones (1999), which are computed as an average of different sub-indicators referring to the degree of trade openness and the quality of the government. Also, see the widely used Economic Freedom of the World index (EFW) or the Global Competitiveness index which are all composite indicators.

Recent empirical and theoretical literature argues how more competition and neck-and-neck markets may incentivize innovation and therefore affect also positively the possibility of adoption of new innovations and technology across countries (Griffith et al, 2006).

Since the IM variables we use in this paper are given at the national level, without sectoral variation, we avoid the sub-index i in them.

The denominator is measured following the standard definition used by the EU to measure the share of innovation within countries or NACE. Innovative firms are those which innovate in product and/or process, including “ongoing or abandoned innovation activities” (process or product).

Due to the fact that the data for the EFW do not precisely coincide in time with those of the CIS database we decided to use in the regressions the data of EFW for the year 1995. This is also due to the fact that some time lag may be experienced between the application of the regulations and their effect on the channels and on innovation adoption.

Conway et al., (2005), “Product Market Regulation in OECD Countries, 1998 to 2003”, OECD Economics Department Working Paper, No 419. This paper is available at: http://www.olis.oecd.org/olis/2005doc.nsf/linkto/ECO-WKP(2005)6.

If we look more in detail the index of Freedom to Trade Internationally we can note that this is built as the average of other sub-indices. These are: taxes on international trade (representing the revenues and mean tariff rates applied in each country as well as the standard deviation of these tariffs), regulatory barriers (as the average of hidden import barriers and cost of importing), actual size of trade sector compared to its expected size (derived from gravity analysis), differences between official exchange rates and black-market rate and finally international capital market controls (as the average of an index controlling for the access of citizens to foreign capital markets and vice versa and the restrictions on the freedom of citizens to engage in capital market exchange with foreigners). All these sub-proxies together give the index for the easiness by which trade transactions may take place across countries.

Other specifications as robustness checks have been tried and can be provided upon request. All of them also perform well even if some of the explanatory variables used as robustness checks do not show statistically significant coefficients and are therefore not used in the system of equations later on in the paper.

If we examine this regulatory and administrative opacity index more in depth we can that note this is a composed measure of different sub-indicators. One of the sub-indicators is the “communication and simplification procedure” index which measures the extent by which national governments reduce information and cooperation costs. This sub-indicator is itself the result of the quantification of different aspects related to the easiness with which business operation can be carried out in the EU member states. To give an example, the third item “There are inquiry points where affected or interested foreign parties can get information on the operation and enforcement of regulations” is actually providing a measure of the extent by which national government foster and ease cooperation among national firms/individuals and other firms/individuals in other countries. Another example is the eighth item of the indicator which measures the presence of “a program to review and reduce the number of licenses and permits required by the national government”. The presence of such type of programs is undoubtedly helpful in fostering cooperation across firms making the start-up of business relations much easier not only at the national level but especially for foreigners who do not have to face too many “unknown” procedures in order to start their businesses abroad.

We re-run the baseline specification of the cooperation channel but using as a dependent variable the degree of cooperation between firms within the same country. Results are basically the same. We argue, therefore, that the same IM policies affecting the cooperation across EU countries are likely to affect also the degree of cooperation within firms in the same country, as one would expect.

The first low-level indicator has already been examined in the text in the analysis for the trade channel. The second, instead, is itself a weighted average of three other items, namely “scope of public enterprise sector”, “size of public enterprise sector” and “direct control over business enterprises”.

The proxy is calculated as the ratio value added/(labor costs+ capital costs).

This variable comes from the Internal Market scoreboard, Eurostat.

We run some alternative specifications through the use of some different explanatory variables. Along with “state control” also a proxy for “regulatory barriers”, “barriers to entrepreneurship” seem to explain the competition level across countries. When these same variables are accompanied also to the “outward oriented policy” item of the OECD PMR indicators, again the signs of the coefficients are those expected so that tighter regulations and direct control by the national government on the private business reduce the competition level (increases the value of the indicator). Other combinations of the same regulatory proxies have been tried such as “Government enterprises and investment as a percentage of total investment” (EFW) and they all show qualitatively the same result.

This variable comes from the EFW index and is based on the Political Risk Component I (Law and Order) from the International Country Risk Guide. Source: PRS Group (various issues), International Country Risk Guide.

In the case of EU countries, we may suspect that the first effect (that of innovation creation) may be stronger than the second (innovation adoption) due to the high-tech nature of innovative European firms if compared to other regions in the world where countries rely on innovation adoption more than in innovation creation. Further econometric investigation is anyway needed to confirm this hypothesis.

Similar results are found in Manca (2010) on a larger cross section of countries for the period 1970–2000.

Additional control variables have also been inserted in column (i). We inserted the R&D expenditures (in order to capture country and sectoral differences in the innovative effort) and other controls such as the share of firms “belonging to a group”, “receiving public funds” or “having made any organizational change lately”. Even if none of these controls shows up to be statistically significant, our main results on the impact of the transmission channels are maintained with the same statistical significance.

The countries for which data are available at the macro level for the required CIS3 items are: Belgium, Germany, Greece, Spain, France, Italy, Luxembourg, Netherlands, Austria, Portugal, Finland, Sweden, Iceland and Norway.

Concerning the micro dataset, we also face difficulties because micro databases are not harmonised between countries. In particular, this concerns areas such as the data type, the code for non-response, etc. Once more this is likely to constrain some cross-country comparisons and a full exploitation of the data. Moreover on the micro database, some missing or inconsistent observations have led to a reduction of the size of the sample. For example, because of individuals answering ”yes” to the items “product and process innovation” and giving no answer to the item “innovation mainly developed by others or in collaboration with other firms”, 622 observations have been dropped from the sample. So, the data set contains 49,139 firms after cleaning. Deleting missing and inconsistent observations has a very weak influence on the spatial and sectoral characteristics of the sample.

Briefly, this refers to the fact that at the macro level we are not able to verify whether a firm is replying yes to the two specific items (adoption in product and process) of the CIS3 necessary for the construction of the innovation adoption indicator (or only one of them) and, therefore, we may double count such a firm.