Can behavioral interventions optimize self-consumption? Evidence from a field experiment with prosumers in Germany

Aligning prosumers' electricity consumption to the availability of self-generated electricity decreases CO 2 emissions and costs. Nudges are proposed as one behavioral intervention to orchestrate such changes. At the same time, fragmented findings in the literature make it challenging to identify suitable behavioral interventions for specific households and contexts-specifically for optimizing self-consumption. We test three sequentially applied interventions (feedback, benchmark, and default) delivered by digital tools in a field experiment with 111 German households with rooftop-photovoltaics. The experiment design with a control-group, baseline measurements, and high-frequency smart-meter-data allows us to examine the causal effects of each intervention for increasing self-consumption. While feedback and benchmark deliver small self-consumption increases (3-4 percent), the smart changing default leads to a 16 percent increase for active participants. In general, households with controllable electric vehicles show stronger effects than those without. For upscaling behavioral interventions for other prosumers, we recommend interventions that require little interaction and energy literacy because even the self-selected, motivated sample rarely interacted with the digital tools.