• Codes of conduct

Codes of conduct

Social Accountability 8000 Standard (SA 8000) and similar corporate codes are promoted as a necessary tool to improve workplace conditions, especially in nations that lack robust enforcement of regulatory standards.These certification standards and corporate codes represent a new form of voluntary “self-governance” of working conditions in the private sector, and are initiated and implemented by companies, labour unions, and non-governmental activist groups.

The number of organisations becoming certified under SA 8000 and similar codes continues to rise, and additional codes governing working conditions continue to emerge, but it is unclear whether these codes represent substantive efforts to improve working conditions or are merely symbolic efforts that allow organisations to score marketing points and counteract stakeholder pressure. Critics also question the skills and incentives of those charged with monitoring and certifying organisations.

This paper reviews existing evaluations of other private codes governing workplace conditions. It discusses the principles and techniques conforming to best practices for evaluation studies in other domains and suggests they should be incorporated in future evaluations of codes like SA 8000. This would bolster their robustness and enable researchers to make clear causal inferences, distinguishing the performance differences between code adopters and non-adopters before adoption and the performance differences that emerge after adoption. The paper argues that evaluations that meet the highest methodological standards are critical to inform the debates about this new form of private governance, and to highlight opportunities for improvement in their standards and monitoring procedures.

The paper also notes that methodological problems encountered in standard focused and comparison studies of voluntary codes could be dealt with through better research design that includes three fundamental research design principles:

• examine performance over time
• compare the performance of participants to a very similar set of non-participants
• address selection bias using randomised trials wherever possible

The paper recommends three alternative research design approaches that can provide ways to overcome selection bias in evaluations of codes such as SA 8000:

• randomisation
• matching to establish quasi-control groups
• using instrumental variables

The paper concludes that the research design criteria proposed are not easy to implement, but are standard practice in other domains of programme evaluation and can help introduce systematic evidence to address the ongoing debates about the effectiveness of certification under SA 8000 and similar codes.

• Full text of document
• Publisher details
• Send to a friend