This study is the first to use a comprehensive decomposition of the Hicks–Moorsteen TFP index developed by O’Donnell (2010a) to analyse efficiency and productivity changes in a banking context. The paper investigates the efficiency and productivity growth of the Iranian banking industry between 2003 and 2008, encompassing pre- and post-2005-reform years. The advantage of this approach over the popular constant-returns-to-scale Malmquist productivity index is that it is free from any assumptions concerning firms’ returns to scale. We assume that the production technology exhibits variable returns to scale. Our findings show that the banking industry’s technical efficiency level – which had improved between 2003 and 2006 – deteriorated after regulatory changes were introduced in Iran. The results obtained also show that during 2006–2007, the industry’s total factor productivity increased by 32 per cent. However, the industry experienced its highest negative scale efficiency rate of 38 per cent and its highest negative efficiency growth of 43 per cent during this period. The industry also witnessed a strong drop in productivity in 2007–2008. Overall, our findings show that while government regulations may have resulted in large advances in the production possibilities set and therefore banks’ productivity over time, the state regulatory changes exacerbated the industry’s scale inefficiencies.