Applying artificial intelligence and quantitative finance for a successful heat transition in the building sector

Counteracting global warming requires intensifying decarbonization efforts across all sectors. To this end, the global residential building sector faces an urgent need to progress towards the climate goals, as it accounts for over a sixth of greenhouse gas emissions and over a quarter of energy consumption, most of which are caused by warm water and space heating and cooling. However, many economically and ecologically sensible retrofit measures are not conducted, among other reasons because of high perceived uncertainty regarding financial savings. Against this background, this doctoral thesis aims to contribute to successfully shaping the heat transition in the residential building sector by investigating three main aspects. The first aspect deals with reducing the perceived risk for energetic retrofitting by providing reliable data-driven decision support, as there is currently a research gap regarding long-term (i. e. , annual) prediction for residential buildings and the resulting consequences of increased prediction accuracy. The findings in this thesis provide strong evidence that data-driven energy quantification methods reduce prediction errors by about 50% compared to the legally prescribed engineering methods. Assuming rational decision-making and setting up an agent-based building stock model, this increase in prediction accuracy translates into a substantial rise in energetic retrofitting from about 0. 98% to 1. 68%. Within the model setting, further prediction accuracy gains allow the retrofit rate to eventually exceed the envisaged 2% to successfully shape the heat transition in the residential building sector. The second aspect deals with understanding and managing the remaining risks connected to energetic retrofitting applying concepts from quantitative finance. To this end, this thesis follows literature and differentiates technological and operational risks (first aspect) from contextual and economic risks (second aspect). The findings indicate that risk perception is crucial for evaluating energetic retrofitting. Moreover, the findings provide the theoretical basis and highlight the potential of diversifying and hedging the remaining risk on the financial markets via risk transfer contracts. The third aspect deals with carefully tailored policy measures by constructing spatially and temporally differentiated incentive mechanisms to allocate scarce financial resources efficiently, maximizing the greenhouse gas emission reductions per monetary unit invested. The findings indicate significant influence from regionally differing socio-economic factors on energetic retrofitting. Moreover, time-dependent subsidy schemes incentivizing early retrofitting reduce greenhouse gas emissions substantially. Assuming rational decision-making, greenhouse gas emission reductions per monetary unit invested for time-dependent subsidy schemes exceed the reductions by static subsidy schemes by up to 675%. In summary, this cumulative doctoral thesis comprises seven research articles and aims to contribute to the heat transition in the residential building sector by applying artificial intelligence and concepts from quantitative finance and deriving managerial and policy implications for all focal aspects.