Evidence of 'modest' impact of codes on pay, performance

What's the point of voluntary codes of conduct? It's not an easy question to answer, given the complexity of how people act and work in organisations. At the top of companies, where power and egos play large roles, the interactions are likely to be multifaceted and hard to measure. But with all the attention on codes of corporate governance, it's worth asking what evidence is there that codes solve the "agency problem" they were intended to address? In a paper for the European Corporate Governance Institute, two scholars from Tilburg in the Netherlands and Exeter in the UK examined data from before the Cadbury reforms in 1992 and after to see what happened to two measures of constraint on the freedom of top managers. They looked at CEO turnover and executive pay plans to see how sensitive they were to performance measures. If companies do poorly, you'd expect to see non-voluntary change at the top and constraint on pay levels. The implementation of the codes, with enhanced disclosure of executive pay and new board structures, didn't seem to influence whether a CEO stayed in post during bad times. Nor did the ownership structure: Companies with large shareholders didn't change poorly performing CEOs any more than those with dispersed shareholding. With regard to CEO remuneration, they sketch a nuanced picture. "[T]he regulatory effort undertaken in the UK over the 1990s has had at best a moderate effect on increasing executives' accountability and performance sensitivity of their turnover," they write. So much for codes.

Source document: The working paper "Managerial Remuneration and Disciplining in the UK: A Tale of Two Governance Regimes," by Luc Renneboog of Tilburg University and Grzegorz Trojanowski of the University of Exeter, is a 52-page pdf file.