Published 23 March 2016

How randomised control trials develop growth policy

– lessons learnt and suggestions for application

The Swedish public support to enterprises amounts to SEK 30 billion. Growth Analysis has reported of the supports unclear effects. This report describes the advantages of the of so called randomised control trials (RCT) in the realm of growth policy. The application of RCT:s will produce clear cut evidence of the results of support instrument as well as benefit the design thereof.

Benefits of randomised control studies

A randomised control trial is a method with high credibility when it comes to identifying the effects of using economic policy instruments as selective aid. In brief, the method means that an intervention such as business aid is distributed randomly among a selected number of units such as companies or persons. The random distribution means that the differences that arise between the aid recipients and the control group can be linked to the intervention.

Because it is easy to link cause and effect in randomised control trials, they have become increasingly popular in several areas of policy, not just in the field of health policy, for example, labour market policy instruments. Lessons learnt from field experiments in development economics are now having an impact on the various forms of support used in aid policy.

However, randomised control trials require that economic policy instruments can formulate specific goals. If that is the case, the method can also facilitate the development of the ”effect logic” of the aid instrument. Randomised control trials may mean that aid policy instruments can be evaluated for precisely the dimension that is most expected to have an effect.

RCT:s demands ethical caution

When randomised control trials are used, certain ethical and judicial circumstances must be made clear and adequately handled. The ethical issues are primarily a) that participation in a randomised trial is voluntary and b) that choosing to decline from participation must not have any economic consequences. The judicial aspects are primarily about how to handle the fact that not all companies that apply for aid can be given aid. The report states that neither the ethical nor the judicial difficulties are such that the method cannot be used.

Recommendations

Growth Analysis concludes that the benefits of RCT:s can be gained for support instru­ments where the support measure is relatively small and the targeted population of enterprises relatively large. The increased utilization of EU de minimini regulation has made such support instruments more common. The report suggests how RCT:s can be applied within the different kinds of voucher schemes initiated by the agencies Vinnova and Tillväxtverket (Agency for enterprise and regional growth).

As with other methods, one must ensure that the advantages of the method are utilised to the full. Randomised trials imply increased costs when gathering data. However, the report claims that these costs must be balanced against the costs that arise when aid policy instru­ments do not result in the expected effects.

Business aid that has no effect can be viewed as wasted tax revenue that amounts to three times the amount of aid. From this perspective, the cost of randomised trials should be seen as a minor problem.

Growth Analysis concludes that randomised control trials can certainly be used in the areas where aid policy instruments are relatively small but targeted at large populations. The increase in the use of the EU’s de minimis regulation for business aid has had the effect that precisely this type of aid has become more common. Growth Analysis suggests also that adequate guidelines and ethical principles are formulated to facilitate the future appli­cations of randomised control trials targeting enterprises.

Title
How randomised control trials develop growth policy – lessons learnt and suggestions for application

Serial number
PM 2016:03

Reference number
2015/265

Download the report in SwedishPDF