Improvements in energy efficiency can help facing the on-going climate and energy crises, yet the energy intensity of economic activities at the global level in recent years has decreased more slowly than it is required to achieve climate goals. Based on this premise, the paper builds a macroeconomic agent-based K+S model to study the effects of different policies on energy efficiency. In the model, energy efficiency of capital goods improves as the outcome of endogenous, bottom-up technical change. Public policies analysed range from indirect policies based on taxes, incentives, and subsidies, rooted in the traditional role of the State as fixing market failures, to direct technological policies, akin to the entrepreneurial state approach, in which a public research laboratory invests in R&D with the aim to establish a new technological paradigm on energy efficiency. Simulation results show that while most policies tested are effective in reducing energy intensity, the public research lab is extremely effective in promoting energy efficiency without deteriorating macroeconomic and public finance conditions. The superiority of the national lab policy, however, emerges on a relatively long time-horizon, highlighting the importance of governments that are patient enough to wait for the returns of that policy and the necessity to complement this strategy with more ''ready to use'' indirect measures. Additionally, results indicate that the macroeconomic rebound effect induced by most of the policies is rather small. Concerns about macroeconomic rebound effects are, therefore, most likely often overstated.