Linking ICT related Innovation Adoption and Productivity: results from micro-aggregated data versus firm-level data.

E-business systems are increasingly considered as important examples of ICT related innovations embodied in software applications, the adoption of which is essential for capturing the potential fruits of several ICT externalities. For analysing the importance of this type of embodied technological progress several routes are open. One route is to look at the data that can be used. In this paper we apply the same modelling strategy to two different types of data: 1) cross-country-industry micro-aggregated data obtained after applying Distributed Micro data Analysis (DMD) and 2) firm-level data, in this case for the Netherlands. Today, the econometric analysis based on firm-level data is often more advanced and more complicated from an econometric point of view than the analysis on aggregated data. We show that DMD can be extended to enable the estimation of more complicated models that feature recent directions in micro-econometric analysis on firm-level data. Our application concerns the innovative use of E-business systems by firms. Using a rich set of cross-country-industry data constructed and tailored by DMD for this purpose, we analyse the adoption of three E-business systems (Eterprise Resourc Planning, Customer Relationship Management, Supply Chain Management). We investigate the complementarities in joint adoption and the productivity effects of adopting systems simultaneously or in isolation. The same exercise is repeated on firm-level data for the Netherlands. Our example illustrates that international benchmarking with more elaborate models on cross-country-industry panel data is feasible after using DMD to tailor the underlying firm-level data for specific research questions. This is an important result in the light of the restrictions on pooling cross-country micro data due to confidentiality rules. We find that the results are more diverging for the estimation of complementarities at the adoption stage than for the productivity effects of (joint) adoption. This result implies that measurement error and unobservable heterogeneity plays a greater role when explaining adoption pattern at the firm-level than at the aggregate level.