Quantitative methods make it possible to investigate relations of cause and effect on a wide basis. These methods include standardised questionnaires, cost-benefit analyses, scientometric and bibliometric analyses, but also econometric methods. Econometrics enables an analysis of the consequences of political action, while at the same time taking account of other influencing factors on the micro- or macroeconomic level. Admittedly, these models are complex and require a high number of observations (cases) as well as many variables that describe the object of the analysis (properties).
Econometric Impact Analyses of Development Programmes
The goal of an econometric impact analysis of measures aimed at stimulating development within a firm is to quantify the effect these measures have had on the firms activities. One question that can be answered with the help of this type of analysis is as follows: to what extent has the innovative activity taking place within the firm been altered by the public subsidies received? To answer this question, the company receiving the subsidy would have to be observed in two states: with and without the influence of the subsidy. Since it is impossible to observe these two states simultaneously, an approximation of counterfactual situation is made with the help of econometric methods. One possible way to approximate the counterfactual situation is through the use of control groups. This method compares subsidised companies with unsubsidised companies from an appropriate control group.
Modern control-group based approaches take into account the fact that public subsides are not obtained by chance, but through various selection processes. On one hand, different firms act in different ways when it comes to applying for subsides (for example because they have different information about what is on offer by way of public subsides). On the other hand, the process of selection carried out by the funding administrators uses non-stochastic mechanisms (for example the so-called Picking-the-Winners Strategy can be employed, by which only the most promising firms receive public subsidies. Both of these cases lead to a distortion of the selection process, which has to be taken into account before consistent results can be achieved in an impact analysis.
Two widely-used econometric evaluation methods are matching models and selection models. These two approaches differ in their fundamental model assumptions and in the way they construct the counterfactual situation. Matching models estimate the effect of subsidies by aligning each subsidised firm with a twin firm (i.e. a firm with identical or similar characteristics) from a suitable control group of unsubsidised firms. The matching process involves using the unsubsidised twin company to approximate the counterfactual situation for a given subsidised firm. If the only difference between the twin firms and their subsidised equivalents is that the twins have not received any subsides, it can be concluded that any differences in the amount of activity related to innovation are due to the public subsidy.
In selection models, the first step (selection equation) is to estimate the probability of a firm receiving subsides, based on the characteristics of the firm and on market conditions. The second step (structural equation) the innovation-related activity of the subsidised and unsubsidised firms is estimated. As a part of this process, the fact that the subsidised companies were selected is taken into account by the use of so-called selection correction terms, which are based on the probability of subsidisation estimated in the first step. The effect of the subsidy is then calculated as the difference between the average estimated levels of innovation-related activity in the subsidised and unsubsidised firms, as adjusted by the selection correction.
The following PDF-Document (at this time available in german only) gives an overview of the studies and analyses that have made use of econometric methods.
Back to top