The building sector is one of the largest consumers of energy and a major contributor to greenhouse gas emissions. Improving the energy efficiency of buildings is therefore critical to reducing environmental impacts and achieving the climate targets set by the European Green Deal. In this context, energy renovations, such as insulation upgrades and heating system improvements, are widely promoted as a key strategy. However, there remains considerable uncertainty around the actual energy savings achieved in practice, partly due to performance gaps and behavioral rebound effects. This thesis aims to quantify the real-world impact of energy-efficient renovations on residential gas consumption in the Netherlands, focusing on older dwellings constructed before 1980.

Utilizing data from the WoON 2018 housing survey, which includes detailed data on dwellings and households, the study employs a multiple linear regression model to isolate the effect of renovations on energy use. By comparing the energy consumption of similar renovated and non-renovated dwellings using statistical methods, this study leverages real-world data to estimate the extent to which renovations reduce energy consumption. The research addresses the influence of occupant behavior and other contextual factors, aiming to provide robust estimates of energy savings attributable to renovations.

The results show that a one-unit improvement in thermal quality, roughly equivalent to upgrading three energy label classes, is associated with a 16.4% average reduction in annual gas consumption. Interestingly, the effect of renovations increases with higher thermal quality levels, contrary to the expectation of diminishing returns. In contrast, differences in gas consumption among homes rated between label G and D are not statistically significant. A national policy scenario, in which all dwellings are upgraded to label D, is estimated to reduce total residential gas consumption by only 0.8–1.3%, and by 1.4–2.3% within the sample. A robustness check incorporating thermostat temperature reveals signs of a rebound effect, suggesting that improved efficiency may lead to increased usage or comfort gains that offset part of the savings.

These findings have important implications for energy and climate policy. First, they underscore the need for ambitious renovation standards, as moderate improvements may yield limited savings. Second, they raise concerns about the effectiveness of using label D as a regulatory or subsidy threshold. Finally, the results support more nuanced cost-benefit evaluations and the targeting of deeper retrofits in older buildings to achieve meaningful reductions in energy use. Overall, this study contributes empirical evidence to inform the design and evaluation of renovation policies aimed at transitioning the residential sector toward near-zero energy buildings by 2050.